Edit on GitHub

sqlglot.dialects.duckdb

   1from __future__ import annotations
   2
   3import typing as t
   4
   5from sqlglot import exp, generator, parser, tokens, transforms
   6from sqlglot.expressions import DATA_TYPE
   7from sqlglot.dialects.dialect import (
   8    Dialect,
   9    JSON_EXTRACT_TYPE,
  10    NormalizationStrategy,
  11    approx_count_distinct_sql,
  12    arrow_json_extract_sql,
  13    binary_from_function,
  14    bool_xor_sql,
  15    build_default_decimal_type,
  16    date_trunc_to_time,
  17    datestrtodate_sql,
  18    no_datetime_sql,
  19    encode_decode_sql,
  20    build_formatted_time,
  21    inline_array_unless_query,
  22    no_comment_column_constraint_sql,
  23    no_time_sql,
  24    no_timestamp_sql,
  25    pivot_column_names,
  26    rename_func,
  27    strposition_sql,
  28    str_to_time_sql,
  29    timestamptrunc_sql,
  30    timestrtotime_sql,
  31    unit_to_var,
  32    unit_to_str,
  33    sha256_sql,
  34    build_regexp_extract,
  35    explode_to_unnest_sql,
  36    no_make_interval_sql,
  37)
  38from sqlglot.generator import unsupported_args
  39from sqlglot.helper import seq_get
  40from sqlglot.tokens import TokenType
  41from sqlglot.parser import binary_range_parser
  42
  43DATETIME_DELTA = t.Union[
  44    exp.DateAdd, exp.TimeAdd, exp.DatetimeAdd, exp.TsOrDsAdd, exp.DateSub, exp.DatetimeSub
  45]
  46
  47WINDOW_FUNCS_WITH_IGNORE_NULLS = (
  48    exp.FirstValue,
  49    exp.LastValue,
  50    exp.Lag,
  51    exp.Lead,
  52    exp.NthValue,
  53)
  54
  55
  56def _date_delta_sql(self: DuckDB.Generator, expression: DATETIME_DELTA) -> str:
  57    this = expression.this
  58    unit = unit_to_var(expression)
  59    op = (
  60        "+"
  61        if isinstance(expression, (exp.DateAdd, exp.TimeAdd, exp.DatetimeAdd, exp.TsOrDsAdd))
  62        else "-"
  63    )
  64
  65    to_type: t.Optional[DATA_TYPE] = None
  66    if isinstance(expression, exp.TsOrDsAdd):
  67        to_type = expression.return_type
  68    elif this.is_string:
  69        # Cast string literals (i.e function parameters) to the appropriate type for +/- interval to work
  70        to_type = (
  71            exp.DataType.Type.DATETIME
  72            if isinstance(expression, (exp.DatetimeAdd, exp.DatetimeSub))
  73            else exp.DataType.Type.DATE
  74        )
  75
  76    this = exp.cast(this, to_type) if to_type else this
  77
  78    return f"{self.sql(this)} {op} {self.sql(exp.Interval(this=expression.expression, unit=unit))}"
  79
  80
  81# BigQuery -> DuckDB conversion for the DATE function
  82def _date_sql(self: DuckDB.Generator, expression: exp.Date) -> str:
  83    result = f"CAST({self.sql(expression, 'this')} AS DATE)"
  84    zone = self.sql(expression, "zone")
  85
  86    if zone:
  87        date_str = self.func("STRFTIME", result, "'%d/%m/%Y'")
  88        date_str = f"{date_str} || ' ' || {zone}"
  89
  90        # This will create a TIMESTAMP with time zone information
  91        result = self.func("STRPTIME", date_str, "'%d/%m/%Y %Z'")
  92
  93    return result
  94
  95
  96# BigQuery -> DuckDB conversion for the TIME_DIFF function
  97def _timediff_sql(self: DuckDB.Generator, expression: exp.TimeDiff) -> str:
  98    this = exp.cast(expression.this, exp.DataType.Type.TIME)
  99    expr = exp.cast(expression.expression, exp.DataType.Type.TIME)
 100
 101    # Although the 2 dialects share similar signatures, BQ seems to inverse
 102    # the sign of the result so the start/end time operands are flipped
 103    return self.func("DATE_DIFF", unit_to_str(expression), expr, this)
 104
 105
 106@unsupported_args(("expression", "DuckDB's ARRAY_SORT does not support a comparator."))
 107def _array_sort_sql(self: DuckDB.Generator, expression: exp.ArraySort) -> str:
 108    return self.func("ARRAY_SORT", expression.this)
 109
 110
 111def _sort_array_sql(self: DuckDB.Generator, expression: exp.SortArray) -> str:
 112    name = "ARRAY_REVERSE_SORT" if expression.args.get("asc") == exp.false() else "ARRAY_SORT"
 113    return self.func(name, expression.this)
 114
 115
 116def _build_sort_array_desc(args: t.List) -> exp.Expression:
 117    return exp.SortArray(this=seq_get(args, 0), asc=exp.false())
 118
 119
 120def _build_date_diff(args: t.List) -> exp.Expression:
 121    return exp.DateDiff(this=seq_get(args, 2), expression=seq_get(args, 1), unit=seq_get(args, 0))
 122
 123
 124def _build_generate_series(end_exclusive: bool = False) -> t.Callable[[t.List], exp.GenerateSeries]:
 125    def _builder(args: t.List) -> exp.GenerateSeries:
 126        # Check https://duckdb.org/docs/sql/functions/nested.html#range-functions
 127        if len(args) == 1:
 128            # DuckDB uses 0 as a default for the series' start when it's omitted
 129            args.insert(0, exp.Literal.number("0"))
 130
 131        gen_series = exp.GenerateSeries.from_arg_list(args)
 132        gen_series.set("is_end_exclusive", end_exclusive)
 133
 134        return gen_series
 135
 136    return _builder
 137
 138
 139def _build_make_timestamp(args: t.List) -> exp.Expression:
 140    if len(args) == 1:
 141        return exp.UnixToTime(this=seq_get(args, 0), scale=exp.UnixToTime.MICROS)
 142
 143    return exp.TimestampFromParts(
 144        year=seq_get(args, 0),
 145        month=seq_get(args, 1),
 146        day=seq_get(args, 2),
 147        hour=seq_get(args, 3),
 148        min=seq_get(args, 4),
 149        sec=seq_get(args, 5),
 150    )
 151
 152
 153def _struct_sql(self: DuckDB.Generator, expression: exp.Struct) -> str:
 154    args: t.List[str] = []
 155
 156    # BigQuery allows inline construction such as "STRUCT<a STRING, b INTEGER>('str', 1)" which is
 157    # canonicalized to "ROW('str', 1) AS STRUCT(a TEXT, b INT)" in DuckDB
 158    # The transformation to ROW will take place if:
 159    #  1. The STRUCT itself does not have proper fields (key := value) as a "proper" STRUCT would
 160    #  2. A cast to STRUCT / ARRAY of STRUCTs is found
 161    ancestor_cast = expression.find_ancestor(exp.Cast)
 162    is_bq_inline_struct = (
 163        (expression.find(exp.PropertyEQ) is None)
 164        and ancestor_cast
 165        and any(
 166            casted_type.is_type(exp.DataType.Type.STRUCT)
 167            for casted_type in ancestor_cast.find_all(exp.DataType)
 168        )
 169    )
 170
 171    for i, expr in enumerate(expression.expressions):
 172        is_property_eq = isinstance(expr, exp.PropertyEQ)
 173        value = expr.expression if is_property_eq else expr
 174
 175        if is_bq_inline_struct:
 176            args.append(self.sql(value))
 177        else:
 178            key = expr.name if is_property_eq else f"_{i}"
 179            args.append(f"{self.sql(exp.Literal.string(key))}: {self.sql(value)}")
 180
 181    csv_args = ", ".join(args)
 182
 183    return f"ROW({csv_args})" if is_bq_inline_struct else f"{{{csv_args}}}"
 184
 185
 186def _datatype_sql(self: DuckDB.Generator, expression: exp.DataType) -> str:
 187    if expression.is_type("array"):
 188        return f"{self.expressions(expression, flat=True)}[{self.expressions(expression, key='values', flat=True)}]"
 189
 190    # Modifiers are not supported for TIME, [TIME | TIMESTAMP] WITH TIME ZONE
 191    if expression.is_type(
 192        exp.DataType.Type.TIME, exp.DataType.Type.TIMETZ, exp.DataType.Type.TIMESTAMPTZ
 193    ):
 194        return expression.this.value
 195
 196    return self.datatype_sql(expression)
 197
 198
 199def _json_format_sql(self: DuckDB.Generator, expression: exp.JSONFormat) -> str:
 200    sql = self.func("TO_JSON", expression.this, expression.args.get("options"))
 201    return f"CAST({sql} AS TEXT)"
 202
 203
 204def _unix_to_time_sql(self: DuckDB.Generator, expression: exp.UnixToTime) -> str:
 205    scale = expression.args.get("scale")
 206    timestamp = expression.this
 207
 208    if scale in (None, exp.UnixToTime.SECONDS):
 209        return self.func("TO_TIMESTAMP", timestamp)
 210    if scale == exp.UnixToTime.MILLIS:
 211        return self.func("EPOCH_MS", timestamp)
 212    if scale == exp.UnixToTime.MICROS:
 213        return self.func("MAKE_TIMESTAMP", timestamp)
 214
 215    return self.func("TO_TIMESTAMP", exp.Div(this=timestamp, expression=exp.func("POW", 10, scale)))
 216
 217
 218WRAPPED_JSON_EXTRACT_EXPRESSIONS = (exp.Binary, exp.Bracket, exp.In)
 219
 220
 221def _arrow_json_extract_sql(self: DuckDB.Generator, expression: JSON_EXTRACT_TYPE) -> str:
 222    arrow_sql = arrow_json_extract_sql(self, expression)
 223    if not expression.same_parent and isinstance(
 224        expression.parent, WRAPPED_JSON_EXTRACT_EXPRESSIONS
 225    ):
 226        arrow_sql = self.wrap(arrow_sql)
 227    return arrow_sql
 228
 229
 230def _implicit_datetime_cast(
 231    arg: t.Optional[exp.Expression], type: exp.DataType.Type = exp.DataType.Type.DATE
 232) -> t.Optional[exp.Expression]:
 233    return exp.cast(arg, type) if isinstance(arg, exp.Literal) else arg
 234
 235
 236def _date_diff_sql(self: DuckDB.Generator, expression: exp.DateDiff) -> str:
 237    this = _implicit_datetime_cast(expression.this)
 238    expr = _implicit_datetime_cast(expression.expression)
 239
 240    return self.func("DATE_DIFF", unit_to_str(expression), expr, this)
 241
 242
 243def _generate_datetime_array_sql(
 244    self: DuckDB.Generator, expression: t.Union[exp.GenerateDateArray, exp.GenerateTimestampArray]
 245) -> str:
 246    is_generate_date_array = isinstance(expression, exp.GenerateDateArray)
 247
 248    type = exp.DataType.Type.DATE if is_generate_date_array else exp.DataType.Type.TIMESTAMP
 249    start = _implicit_datetime_cast(expression.args.get("start"), type=type)
 250    end = _implicit_datetime_cast(expression.args.get("end"), type=type)
 251
 252    # BQ's GENERATE_DATE_ARRAY & GENERATE_TIMESTAMP_ARRAY are transformed to DuckDB'S GENERATE_SERIES
 253    gen_series: t.Union[exp.GenerateSeries, exp.Cast] = exp.GenerateSeries(
 254        start=start, end=end, step=expression.args.get("step")
 255    )
 256
 257    if is_generate_date_array:
 258        # The GENERATE_SERIES result type is TIMESTAMP array, so to match BQ's semantics for
 259        # GENERATE_DATE_ARRAY we must cast it back to DATE array
 260        gen_series = exp.cast(gen_series, exp.DataType.build("ARRAY<DATE>"))
 261
 262    return self.sql(gen_series)
 263
 264
 265def _json_extract_value_array_sql(
 266    self: DuckDB.Generator, expression: exp.JSONValueArray | exp.JSONExtractArray
 267) -> str:
 268    json_extract = exp.JSONExtract(this=expression.this, expression=expression.expression)
 269    data_type = "ARRAY<STRING>" if isinstance(expression, exp.JSONValueArray) else "ARRAY<JSON>"
 270    return self.sql(exp.cast(json_extract, to=exp.DataType.build(data_type)))
 271
 272
 273class DuckDB(Dialect):
 274    NULL_ORDERING = "nulls_are_last"
 275    SUPPORTS_USER_DEFINED_TYPES = False
 276    SAFE_DIVISION = True
 277    INDEX_OFFSET = 1
 278    CONCAT_COALESCE = True
 279    SUPPORTS_ORDER_BY_ALL = True
 280    SUPPORTS_FIXED_SIZE_ARRAYS = True
 281    STRICT_JSON_PATH_SYNTAX = False
 282
 283    # https://duckdb.org/docs/sql/introduction.html#creating-a-new-table
 284    NORMALIZATION_STRATEGY = NormalizationStrategy.CASE_INSENSITIVE
 285
 286    def to_json_path(self, path: t.Optional[exp.Expression]) -> t.Optional[exp.Expression]:
 287        if isinstance(path, exp.Literal):
 288            # DuckDB also supports the JSON pointer syntax, where every path starts with a `/`.
 289            # Additionally, it allows accessing the back of lists using the `[#-i]` syntax.
 290            # This check ensures we'll avoid trying to parse these as JSON paths, which can
 291            # either result in a noisy warning or in an invalid representation of the path.
 292            path_text = path.name
 293            if path_text.startswith("/") or "[#" in path_text:
 294                return path
 295
 296        return super().to_json_path(path)
 297
 298    class Tokenizer(tokens.Tokenizer):
 299        BYTE_STRINGS = [("e'", "'"), ("E'", "'")]
 300        HEREDOC_STRINGS = ["$"]
 301
 302        HEREDOC_TAG_IS_IDENTIFIER = True
 303        HEREDOC_STRING_ALTERNATIVE = TokenType.PARAMETER
 304
 305        KEYWORDS = {
 306            **tokens.Tokenizer.KEYWORDS,
 307            "//": TokenType.DIV,
 308            "**": TokenType.DSTAR,
 309            "^@": TokenType.CARET_AT,
 310            "@>": TokenType.AT_GT,
 311            "<@": TokenType.LT_AT,
 312            "ATTACH": TokenType.ATTACH,
 313            "BINARY": TokenType.VARBINARY,
 314            "BITSTRING": TokenType.BIT,
 315            "BPCHAR": TokenType.TEXT,
 316            "CHAR": TokenType.TEXT,
 317            "CHARACTER VARYING": TokenType.TEXT,
 318            "DETACH": TokenType.DETACH,
 319            "EXCLUDE": TokenType.EXCEPT,
 320            "LOGICAL": TokenType.BOOLEAN,
 321            "ONLY": TokenType.ONLY,
 322            "PIVOT_WIDER": TokenType.PIVOT,
 323            "POSITIONAL": TokenType.POSITIONAL,
 324            "SIGNED": TokenType.INT,
 325            "STRING": TokenType.TEXT,
 326            "SUMMARIZE": TokenType.SUMMARIZE,
 327            "TIMESTAMP_S": TokenType.TIMESTAMP_S,
 328            "TIMESTAMP_MS": TokenType.TIMESTAMP_MS,
 329            "TIMESTAMP_NS": TokenType.TIMESTAMP_NS,
 330            "TIMESTAMP_US": TokenType.TIMESTAMP,
 331            "UBIGINT": TokenType.UBIGINT,
 332            "UINTEGER": TokenType.UINT,
 333            "USMALLINT": TokenType.USMALLINT,
 334            "UTINYINT": TokenType.UTINYINT,
 335            "VARCHAR": TokenType.TEXT,
 336        }
 337        KEYWORDS.pop("/*+")
 338
 339        SINGLE_TOKENS = {
 340            **tokens.Tokenizer.SINGLE_TOKENS,
 341            "$": TokenType.PARAMETER,
 342        }
 343
 344    class Parser(parser.Parser):
 345        BITWISE = {
 346            **parser.Parser.BITWISE,
 347            TokenType.TILDA: exp.RegexpLike,
 348        }
 349        BITWISE.pop(TokenType.CARET)
 350
 351        RANGE_PARSERS = {
 352            **parser.Parser.RANGE_PARSERS,
 353            TokenType.DAMP: binary_range_parser(exp.ArrayOverlaps),
 354            TokenType.CARET_AT: binary_range_parser(exp.StartsWith),
 355        }
 356
 357        EXPONENT = {
 358            **parser.Parser.EXPONENT,
 359            TokenType.CARET: exp.Pow,
 360            TokenType.DSTAR: exp.Pow,
 361        }
 362
 363        FUNCTIONS_WITH_ALIASED_ARGS = {*parser.Parser.FUNCTIONS_WITH_ALIASED_ARGS, "STRUCT_PACK"}
 364
 365        FUNCTIONS = {
 366            **parser.Parser.FUNCTIONS,
 367            "ARRAY_REVERSE_SORT": _build_sort_array_desc,
 368            "ARRAY_SORT": exp.SortArray.from_arg_list,
 369            "DATEDIFF": _build_date_diff,
 370            "DATE_DIFF": _build_date_diff,
 371            "DATE_TRUNC": date_trunc_to_time,
 372            "DATETRUNC": date_trunc_to_time,
 373            "DECODE": lambda args: exp.Decode(
 374                this=seq_get(args, 0), charset=exp.Literal.string("utf-8")
 375            ),
 376            "ENCODE": lambda args: exp.Encode(
 377                this=seq_get(args, 0), charset=exp.Literal.string("utf-8")
 378            ),
 379            "EPOCH": exp.TimeToUnix.from_arg_list,
 380            "EPOCH_MS": lambda args: exp.UnixToTime(
 381                this=seq_get(args, 0), scale=exp.UnixToTime.MILLIS
 382            ),
 383            "JSON": exp.ParseJSON.from_arg_list,
 384            "JSON_EXTRACT_PATH": parser.build_extract_json_with_path(exp.JSONExtract),
 385            "JSON_EXTRACT_STRING": parser.build_extract_json_with_path(exp.JSONExtractScalar),
 386            "LIST_HAS": exp.ArrayContains.from_arg_list,
 387            "LIST_REVERSE_SORT": _build_sort_array_desc,
 388            "LIST_SORT": exp.SortArray.from_arg_list,
 389            "LIST_VALUE": lambda args: exp.Array(expressions=args),
 390            "MAKE_TIME": exp.TimeFromParts.from_arg_list,
 391            "MAKE_TIMESTAMP": _build_make_timestamp,
 392            "QUANTILE_CONT": exp.PercentileCont.from_arg_list,
 393            "QUANTILE_DISC": exp.PercentileDisc.from_arg_list,
 394            "REGEXP_EXTRACT": build_regexp_extract(exp.RegexpExtract),
 395            "REGEXP_EXTRACT_ALL": build_regexp_extract(exp.RegexpExtractAll),
 396            "REGEXP_MATCHES": exp.RegexpLike.from_arg_list,
 397            "REGEXP_REPLACE": lambda args: exp.RegexpReplace(
 398                this=seq_get(args, 0),
 399                expression=seq_get(args, 1),
 400                replacement=seq_get(args, 2),
 401                modifiers=seq_get(args, 3),
 402            ),
 403            "STRFTIME": build_formatted_time(exp.TimeToStr, "duckdb"),
 404            "STRING_SPLIT": exp.Split.from_arg_list,
 405            "STRING_SPLIT_REGEX": exp.RegexpSplit.from_arg_list,
 406            "STRING_TO_ARRAY": exp.Split.from_arg_list,
 407            "STRPTIME": build_formatted_time(exp.StrToTime, "duckdb"),
 408            "STRUCT_PACK": exp.Struct.from_arg_list,
 409            "STR_SPLIT": exp.Split.from_arg_list,
 410            "STR_SPLIT_REGEX": exp.RegexpSplit.from_arg_list,
 411            "TO_TIMESTAMP": exp.UnixToTime.from_arg_list,
 412            "UNNEST": exp.Explode.from_arg_list,
 413            "XOR": binary_from_function(exp.BitwiseXor),
 414            "GENERATE_SERIES": _build_generate_series(),
 415            "RANGE": _build_generate_series(end_exclusive=True),
 416            "EDITDIST3": exp.Levenshtein.from_arg_list,
 417        }
 418
 419        FUNCTIONS.pop("DATE_SUB")
 420        FUNCTIONS.pop("GLOB")
 421
 422        FUNCTION_PARSERS = parser.Parser.FUNCTION_PARSERS.copy()
 423        FUNCTION_PARSERS.pop("DECODE")
 424
 425        NO_PAREN_FUNCTION_PARSERS = {
 426            **parser.Parser.NO_PAREN_FUNCTION_PARSERS,
 427            "MAP": lambda self: self._parse_map(),
 428        }
 429
 430        TABLE_ALIAS_TOKENS = parser.Parser.TABLE_ALIAS_TOKENS - {
 431            TokenType.SEMI,
 432            TokenType.ANTI,
 433        }
 434
 435        PLACEHOLDER_PARSERS = {
 436            **parser.Parser.PLACEHOLDER_PARSERS,
 437            TokenType.PARAMETER: lambda self: (
 438                self.expression(exp.Placeholder, this=self._prev.text)
 439                if self._match(TokenType.NUMBER) or self._match_set(self.ID_VAR_TOKENS)
 440                else None
 441            ),
 442        }
 443
 444        TYPE_CONVERTERS = {
 445            # https://duckdb.org/docs/sql/data_types/numeric
 446            exp.DataType.Type.DECIMAL: build_default_decimal_type(precision=18, scale=3),
 447            # https://duckdb.org/docs/sql/data_types/text
 448            exp.DataType.Type.TEXT: lambda dtype: exp.DataType.build("TEXT"),
 449        }
 450
 451        STATEMENT_PARSERS = {
 452            **parser.Parser.STATEMENT_PARSERS,
 453            TokenType.ATTACH: lambda self: self._parse_attach_detach(),
 454            TokenType.DETACH: lambda self: self._parse_attach_detach(is_attach=False),
 455        }
 456
 457        def _parse_table_sample(self, as_modifier: bool = False) -> t.Optional[exp.TableSample]:
 458            # https://duckdb.org/docs/sql/samples.html
 459            sample = super()._parse_table_sample(as_modifier=as_modifier)
 460            if sample and not sample.args.get("method"):
 461                if sample.args.get("size"):
 462                    sample.set("method", exp.var("RESERVOIR"))
 463                else:
 464                    sample.set("method", exp.var("SYSTEM"))
 465
 466            return sample
 467
 468        def _parse_bracket(
 469            self, this: t.Optional[exp.Expression] = None
 470        ) -> t.Optional[exp.Expression]:
 471            bracket = super()._parse_bracket(this)
 472            if isinstance(bracket, exp.Bracket):
 473                bracket.set("returns_list_for_maps", True)
 474
 475            return bracket
 476
 477        def _parse_map(self) -> exp.ToMap | exp.Map:
 478            if self._match(TokenType.L_BRACE, advance=False):
 479                return self.expression(exp.ToMap, this=self._parse_bracket())
 480
 481            args = self._parse_wrapped_csv(self._parse_assignment)
 482            return self.expression(exp.Map, keys=seq_get(args, 0), values=seq_get(args, 1))
 483
 484        def _parse_struct_types(self, type_required: bool = False) -> t.Optional[exp.Expression]:
 485            return self._parse_field_def()
 486
 487        def _pivot_column_names(self, aggregations: t.List[exp.Expression]) -> t.List[str]:
 488            if len(aggregations) == 1:
 489                return super()._pivot_column_names(aggregations)
 490            return pivot_column_names(aggregations, dialect="duckdb")
 491
 492        def _parse_attach_detach(self, is_attach=True) -> exp.Attach | exp.Detach:
 493            def _parse_attach_option() -> exp.AttachOption:
 494                return self.expression(
 495                    exp.AttachOption,
 496                    this=self._parse_var(any_token=True),
 497                    expression=self._parse_field(any_token=True),
 498                )
 499
 500            self._match(TokenType.DATABASE)
 501            exists = self._parse_exists(not_=is_attach)
 502            this = self._parse_alias(self._parse_primary_or_var(), explicit=True)
 503
 504            if self._match(TokenType.L_PAREN, advance=False):
 505                expressions = self._parse_wrapped_csv(_parse_attach_option)
 506            else:
 507                expressions = None
 508
 509            return (
 510                self.expression(exp.Attach, this=this, exists=exists, expressions=expressions)
 511                if is_attach
 512                else self.expression(exp.Detach, this=this, exists=exists)
 513            )
 514
 515    class Generator(generator.Generator):
 516        PARAMETER_TOKEN = "$"
 517        NAMED_PLACEHOLDER_TOKEN = "$"
 518        JOIN_HINTS = False
 519        TABLE_HINTS = False
 520        QUERY_HINTS = False
 521        LIMIT_FETCH = "LIMIT"
 522        STRUCT_DELIMITER = ("(", ")")
 523        RENAME_TABLE_WITH_DB = False
 524        NVL2_SUPPORTED = False
 525        SEMI_ANTI_JOIN_WITH_SIDE = False
 526        TABLESAMPLE_KEYWORDS = "USING SAMPLE"
 527        TABLESAMPLE_SEED_KEYWORD = "REPEATABLE"
 528        LAST_DAY_SUPPORTS_DATE_PART = False
 529        JSON_KEY_VALUE_PAIR_SEP = ","
 530        IGNORE_NULLS_IN_FUNC = True
 531        JSON_PATH_BRACKETED_KEY_SUPPORTED = False
 532        SUPPORTS_CREATE_TABLE_LIKE = False
 533        MULTI_ARG_DISTINCT = False
 534        CAN_IMPLEMENT_ARRAY_ANY = True
 535        SUPPORTS_TO_NUMBER = False
 536        COPY_HAS_INTO_KEYWORD = False
 537        STAR_EXCEPT = "EXCLUDE"
 538        PAD_FILL_PATTERN_IS_REQUIRED = True
 539        ARRAY_CONCAT_IS_VAR_LEN = False
 540        ARRAY_SIZE_DIM_REQUIRED = False
 541
 542        TRANSFORMS = {
 543            **generator.Generator.TRANSFORMS,
 544            exp.ApproxDistinct: approx_count_distinct_sql,
 545            exp.Array: inline_array_unless_query,
 546            exp.ArrayFilter: rename_func("LIST_FILTER"),
 547            exp.ArraySort: _array_sort_sql,
 548            exp.ArraySum: rename_func("LIST_SUM"),
 549            exp.BitwiseXor: rename_func("XOR"),
 550            exp.CommentColumnConstraint: no_comment_column_constraint_sql,
 551            exp.CurrentDate: lambda *_: "CURRENT_DATE",
 552            exp.CurrentTime: lambda *_: "CURRENT_TIME",
 553            exp.CurrentTimestamp: lambda *_: "CURRENT_TIMESTAMP",
 554            exp.DayOfMonth: rename_func("DAYOFMONTH"),
 555            exp.DayOfWeek: rename_func("DAYOFWEEK"),
 556            exp.DayOfWeekIso: rename_func("ISODOW"),
 557            exp.DayOfYear: rename_func("DAYOFYEAR"),
 558            exp.DataType: _datatype_sql,
 559            exp.Date: _date_sql,
 560            exp.DateAdd: _date_delta_sql,
 561            exp.DateFromParts: rename_func("MAKE_DATE"),
 562            exp.DateSub: _date_delta_sql,
 563            exp.DateDiff: _date_diff_sql,
 564            exp.DateStrToDate: datestrtodate_sql,
 565            exp.Datetime: no_datetime_sql,
 566            exp.DatetimeSub: _date_delta_sql,
 567            exp.DatetimeAdd: _date_delta_sql,
 568            exp.DateToDi: lambda self,
 569            e: f"CAST(STRFTIME({self.sql(e, 'this')}, {DuckDB.DATEINT_FORMAT}) AS INT)",
 570            exp.Decode: lambda self, e: encode_decode_sql(self, e, "DECODE", replace=False),
 571            exp.DiToDate: lambda self,
 572            e: f"CAST(STRPTIME(CAST({self.sql(e, 'this')} AS TEXT), {DuckDB.DATEINT_FORMAT}) AS DATE)",
 573            exp.Encode: lambda self, e: encode_decode_sql(self, e, "ENCODE", replace=False),
 574            exp.GenerateDateArray: _generate_datetime_array_sql,
 575            exp.GenerateTimestampArray: _generate_datetime_array_sql,
 576            exp.Explode: rename_func("UNNEST"),
 577            exp.IntDiv: lambda self, e: self.binary(e, "//"),
 578            exp.IsInf: rename_func("ISINF"),
 579            exp.IsNan: rename_func("ISNAN"),
 580            exp.JSONBExists: rename_func("JSON_EXISTS"),
 581            exp.JSONExtract: _arrow_json_extract_sql,
 582            exp.JSONExtractArray: _json_extract_value_array_sql,
 583            exp.JSONExtractScalar: _arrow_json_extract_sql,
 584            exp.JSONFormat: _json_format_sql,
 585            exp.JSONValueArray: _json_extract_value_array_sql,
 586            exp.Lateral: explode_to_unnest_sql,
 587            exp.LogicalOr: rename_func("BOOL_OR"),
 588            exp.LogicalAnd: rename_func("BOOL_AND"),
 589            exp.MakeInterval: lambda self, e: no_make_interval_sql(self, e, sep=" "),
 590            exp.MD5Digest: lambda self, e: self.func("UNHEX", self.func("MD5", e.this)),
 591            exp.MonthsBetween: lambda self, e: self.func(
 592                "DATEDIFF",
 593                "'month'",
 594                exp.cast(e.expression, exp.DataType.Type.TIMESTAMP, copy=True),
 595                exp.cast(e.this, exp.DataType.Type.TIMESTAMP, copy=True),
 596            ),
 597            exp.PercentileCont: rename_func("QUANTILE_CONT"),
 598            exp.PercentileDisc: rename_func("QUANTILE_DISC"),
 599            # DuckDB doesn't allow qualified columns inside of PIVOT expressions.
 600            # See: https://github.com/duckdb/duckdb/blob/671faf92411182f81dce42ac43de8bfb05d9909e/src/planner/binder/tableref/bind_pivot.cpp#L61-L62
 601            exp.Pivot: transforms.preprocess([transforms.unqualify_columns]),
 602            exp.RegexpReplace: lambda self, e: self.func(
 603                "REGEXP_REPLACE",
 604                e.this,
 605                e.expression,
 606                e.args.get("replacement"),
 607                e.args.get("modifiers"),
 608            ),
 609            exp.RegexpLike: rename_func("REGEXP_MATCHES"),
 610            exp.RegexpILike: lambda self, e: self.func(
 611                "REGEXP_MATCHES", e.this, e.expression, exp.Literal.string("i")
 612            ),
 613            exp.RegexpSplit: rename_func("STR_SPLIT_REGEX"),
 614            exp.Return: lambda self, e: self.sql(e, "this"),
 615            exp.ReturnsProperty: lambda self, e: "TABLE" if isinstance(e.this, exp.Schema) else "",
 616            exp.Rand: rename_func("RANDOM"),
 617            exp.SHA: rename_func("SHA1"),
 618            exp.SHA2: sha256_sql,
 619            exp.Split: rename_func("STR_SPLIT"),
 620            exp.SortArray: _sort_array_sql,
 621            exp.StrPosition: strposition_sql,
 622            exp.StrToUnix: lambda self, e: self.func(
 623                "EPOCH", self.func("STRPTIME", e.this, self.format_time(e))
 624            ),
 625            exp.Struct: _struct_sql,
 626            exp.Transform: rename_func("LIST_TRANSFORM"),
 627            exp.TimeAdd: _date_delta_sql,
 628            exp.Time: no_time_sql,
 629            exp.TimeDiff: _timediff_sql,
 630            exp.Timestamp: no_timestamp_sql,
 631            exp.TimestampDiff: lambda self, e: self.func(
 632                "DATE_DIFF", exp.Literal.string(e.unit), e.expression, e.this
 633            ),
 634            exp.TimestampTrunc: timestamptrunc_sql(),
 635            exp.TimeStrToDate: lambda self, e: self.sql(exp.cast(e.this, exp.DataType.Type.DATE)),
 636            exp.TimeStrToTime: timestrtotime_sql,
 637            exp.TimeStrToUnix: lambda self, e: self.func(
 638                "EPOCH", exp.cast(e.this, exp.DataType.Type.TIMESTAMP)
 639            ),
 640            exp.TimeToStr: lambda self, e: self.func("STRFTIME", e.this, self.format_time(e)),
 641            exp.TimeToUnix: rename_func("EPOCH"),
 642            exp.TsOrDiToDi: lambda self,
 643            e: f"CAST(SUBSTR(REPLACE(CAST({self.sql(e, 'this')} AS TEXT), '-', ''), 1, 8) AS INT)",
 644            exp.TsOrDsAdd: _date_delta_sql,
 645            exp.TsOrDsDiff: lambda self, e: self.func(
 646                "DATE_DIFF",
 647                f"'{e.args.get('unit') or 'DAY'}'",
 648                exp.cast(e.expression, exp.DataType.Type.TIMESTAMP),
 649                exp.cast(e.this, exp.DataType.Type.TIMESTAMP),
 650            ),
 651            exp.UnixToStr: lambda self, e: self.func(
 652                "STRFTIME", self.func("TO_TIMESTAMP", e.this), self.format_time(e)
 653            ),
 654            exp.DatetimeTrunc: lambda self, e: self.func(
 655                "DATE_TRUNC", unit_to_str(e), exp.cast(e.this, exp.DataType.Type.DATETIME)
 656            ),
 657            exp.UnixToTime: _unix_to_time_sql,
 658            exp.UnixToTimeStr: lambda self, e: f"CAST(TO_TIMESTAMP({self.sql(e, 'this')}) AS TEXT)",
 659            exp.VariancePop: rename_func("VAR_POP"),
 660            exp.WeekOfYear: rename_func("WEEKOFYEAR"),
 661            exp.Xor: bool_xor_sql,
 662            exp.Levenshtein: unsupported_args("ins_cost", "del_cost", "sub_cost", "max_dist")(
 663                rename_func("LEVENSHTEIN")
 664            ),
 665        }
 666
 667        SUPPORTED_JSON_PATH_PARTS = {
 668            exp.JSONPathKey,
 669            exp.JSONPathRoot,
 670            exp.JSONPathSubscript,
 671            exp.JSONPathWildcard,
 672        }
 673
 674        TYPE_MAPPING = {
 675            **generator.Generator.TYPE_MAPPING,
 676            exp.DataType.Type.BINARY: "BLOB",
 677            exp.DataType.Type.BPCHAR: "TEXT",
 678            exp.DataType.Type.CHAR: "TEXT",
 679            exp.DataType.Type.DATETIME: "TIMESTAMP",
 680            exp.DataType.Type.FLOAT: "REAL",
 681            exp.DataType.Type.NCHAR: "TEXT",
 682            exp.DataType.Type.NVARCHAR: "TEXT",
 683            exp.DataType.Type.UINT: "UINTEGER",
 684            exp.DataType.Type.VARBINARY: "BLOB",
 685            exp.DataType.Type.ROWVERSION: "BLOB",
 686            exp.DataType.Type.VARCHAR: "TEXT",
 687            exp.DataType.Type.TIMESTAMPNTZ: "TIMESTAMP",
 688            exp.DataType.Type.TIMESTAMP_S: "TIMESTAMP_S",
 689            exp.DataType.Type.TIMESTAMP_MS: "TIMESTAMP_MS",
 690            exp.DataType.Type.TIMESTAMP_NS: "TIMESTAMP_NS",
 691        }
 692
 693        # https://github.com/duckdb/duckdb/blob/ff7f24fd8e3128d94371827523dae85ebaf58713/third_party/libpg_query/grammar/keywords/reserved_keywords.list#L1-L77
 694        RESERVED_KEYWORDS = {
 695            "array",
 696            "analyse",
 697            "union",
 698            "all",
 699            "when",
 700            "in_p",
 701            "default",
 702            "create_p",
 703            "window",
 704            "asymmetric",
 705            "to",
 706            "else",
 707            "localtime",
 708            "from",
 709            "end_p",
 710            "select",
 711            "current_date",
 712            "foreign",
 713            "with",
 714            "grant",
 715            "session_user",
 716            "or",
 717            "except",
 718            "references",
 719            "fetch",
 720            "limit",
 721            "group_p",
 722            "leading",
 723            "into",
 724            "collate",
 725            "offset",
 726            "do",
 727            "then",
 728            "localtimestamp",
 729            "check_p",
 730            "lateral_p",
 731            "current_role",
 732            "where",
 733            "asc_p",
 734            "placing",
 735            "desc_p",
 736            "user",
 737            "unique",
 738            "initially",
 739            "column",
 740            "both",
 741            "some",
 742            "as",
 743            "any",
 744            "only",
 745            "deferrable",
 746            "null_p",
 747            "current_time",
 748            "true_p",
 749            "table",
 750            "case",
 751            "trailing",
 752            "variadic",
 753            "for",
 754            "on",
 755            "distinct",
 756            "false_p",
 757            "not",
 758            "constraint",
 759            "current_timestamp",
 760            "returning",
 761            "primary",
 762            "intersect",
 763            "having",
 764            "analyze",
 765            "current_user",
 766            "and",
 767            "cast",
 768            "symmetric",
 769            "using",
 770            "order",
 771            "current_catalog",
 772        }
 773
 774        UNWRAPPED_INTERVAL_VALUES = (exp.Literal, exp.Paren)
 775
 776        # DuckDB doesn't generally support CREATE TABLE .. properties
 777        # https://duckdb.org/docs/sql/statements/create_table.html
 778        PROPERTIES_LOCATION = {
 779            prop: exp.Properties.Location.UNSUPPORTED
 780            for prop in generator.Generator.PROPERTIES_LOCATION
 781        }
 782
 783        # There are a few exceptions (e.g. temporary tables) which are supported or
 784        # can be transpiled to DuckDB, so we explicitly override them accordingly
 785        PROPERTIES_LOCATION[exp.LikeProperty] = exp.Properties.Location.POST_SCHEMA
 786        PROPERTIES_LOCATION[exp.TemporaryProperty] = exp.Properties.Location.POST_CREATE
 787        PROPERTIES_LOCATION[exp.ReturnsProperty] = exp.Properties.Location.POST_ALIAS
 788
 789        def fromiso8601timestamp_sql(self, expression: exp.FromISO8601Timestamp) -> str:
 790            return self.sql(exp.cast(expression.this, exp.DataType.Type.TIMESTAMPTZ))
 791
 792        def strtotime_sql(self, expression: exp.StrToTime) -> str:
 793            if expression.args.get("safe"):
 794                formatted_time = self.format_time(expression)
 795                return f"CAST({self.func('TRY_STRPTIME', expression.this, formatted_time)} AS TIMESTAMP)"
 796            return str_to_time_sql(self, expression)
 797
 798        def strtodate_sql(self, expression: exp.StrToDate) -> str:
 799            if expression.args.get("safe"):
 800                formatted_time = self.format_time(expression)
 801                return f"CAST({self.func('TRY_STRPTIME', expression.this, formatted_time)} AS DATE)"
 802            return f"CAST({str_to_time_sql(self, expression)} AS DATE)"
 803
 804        def parsejson_sql(self, expression: exp.ParseJSON) -> str:
 805            arg = expression.this
 806            if expression.args.get("safe"):
 807                return self.sql(exp.case().when(exp.func("json_valid", arg), arg).else_(exp.null()))
 808            return self.func("JSON", arg)
 809
 810        def timefromparts_sql(self, expression: exp.TimeFromParts) -> str:
 811            nano = expression.args.get("nano")
 812            if nano is not None:
 813                expression.set(
 814                    "sec", expression.args["sec"] + nano.pop() / exp.Literal.number(1000000000.0)
 815                )
 816
 817            return rename_func("MAKE_TIME")(self, expression)
 818
 819        def timestampfromparts_sql(self, expression: exp.TimestampFromParts) -> str:
 820            sec = expression.args["sec"]
 821
 822            milli = expression.args.get("milli")
 823            if milli is not None:
 824                sec += milli.pop() / exp.Literal.number(1000.0)
 825
 826            nano = expression.args.get("nano")
 827            if nano is not None:
 828                sec += nano.pop() / exp.Literal.number(1000000000.0)
 829
 830            if milli or nano:
 831                expression.set("sec", sec)
 832
 833            return rename_func("MAKE_TIMESTAMP")(self, expression)
 834
 835        def tablesample_sql(
 836            self,
 837            expression: exp.TableSample,
 838            tablesample_keyword: t.Optional[str] = None,
 839        ) -> str:
 840            if not isinstance(expression.parent, exp.Select):
 841                # This sample clause only applies to a single source, not the entire resulting relation
 842                tablesample_keyword = "TABLESAMPLE"
 843
 844            if expression.args.get("size"):
 845                method = expression.args.get("method")
 846                if method and method.name.upper() != "RESERVOIR":
 847                    self.unsupported(
 848                        f"Sampling method {method} is not supported with a discrete sample count, "
 849                        "defaulting to reservoir sampling"
 850                    )
 851                    expression.set("method", exp.var("RESERVOIR"))
 852
 853            return super().tablesample_sql(expression, tablesample_keyword=tablesample_keyword)
 854
 855        def interval_sql(self, expression: exp.Interval) -> str:
 856            multiplier: t.Optional[int] = None
 857            unit = expression.text("unit").lower()
 858
 859            if unit.startswith("week"):
 860                multiplier = 7
 861            if unit.startswith("quarter"):
 862                multiplier = 90
 863
 864            if multiplier:
 865                return f"({multiplier} * {super().interval_sql(exp.Interval(this=expression.this, unit=exp.var('DAY')))})"
 866
 867            return super().interval_sql(expression)
 868
 869        def columndef_sql(self, expression: exp.ColumnDef, sep: str = " ") -> str:
 870            if isinstance(expression.parent, exp.UserDefinedFunction):
 871                return self.sql(expression, "this")
 872            return super().columndef_sql(expression, sep)
 873
 874        def join_sql(self, expression: exp.Join) -> str:
 875            if (
 876                expression.side == "LEFT"
 877                and not expression.args.get("on")
 878                and isinstance(expression.this, exp.Unnest)
 879            ):
 880                # Some dialects support `LEFT JOIN UNNEST(...)` without an explicit ON clause
 881                # DuckDB doesn't, but we can just add a dummy ON clause that is always true
 882                return super().join_sql(expression.on(exp.true()))
 883
 884            return super().join_sql(expression)
 885
 886        def generateseries_sql(self, expression: exp.GenerateSeries) -> str:
 887            # GENERATE_SERIES(a, b) -> [a, b], RANGE(a, b) -> [a, b)
 888            if expression.args.get("is_end_exclusive"):
 889                return rename_func("RANGE")(self, expression)
 890
 891            return self.function_fallback_sql(expression)
 892
 893        def bracket_sql(self, expression: exp.Bracket) -> str:
 894            this = expression.this
 895            if isinstance(this, exp.Array):
 896                this.replace(exp.paren(this))
 897
 898            bracket = super().bracket_sql(expression)
 899
 900            if not expression.args.get("returns_list_for_maps"):
 901                if not this.type:
 902                    from sqlglot.optimizer.annotate_types import annotate_types
 903
 904                    this = annotate_types(this)
 905
 906                if this.is_type(exp.DataType.Type.MAP):
 907                    bracket = f"({bracket})[1]"
 908
 909            return bracket
 910
 911        def withingroup_sql(self, expression: exp.WithinGroup) -> str:
 912            expression_sql = self.sql(expression, "expression")
 913
 914            func = expression.this
 915            if isinstance(func, exp.PERCENTILES):
 916                # Make the order key the first arg and slide the fraction to the right
 917                # https://duckdb.org/docs/sql/aggregates#ordered-set-aggregate-functions
 918                order_col = expression.find(exp.Ordered)
 919                if order_col:
 920                    func.set("expression", func.this)
 921                    func.set("this", order_col.this)
 922
 923            this = self.sql(expression, "this").rstrip(")")
 924
 925            return f"{this}{expression_sql})"
 926
 927        def length_sql(self, expression: exp.Length) -> str:
 928            arg = expression.this
 929
 930            # Dialects like BQ and Snowflake also accept binary values as args, so
 931            # DDB will attempt to infer the type or resort to case/when resolution
 932            if not expression.args.get("binary") or arg.is_string:
 933                return self.func("LENGTH", arg)
 934
 935            if not arg.type:
 936                from sqlglot.optimizer.annotate_types import annotate_types
 937
 938                arg = annotate_types(arg)
 939
 940            if arg.is_type(*exp.DataType.TEXT_TYPES):
 941                return self.func("LENGTH", arg)
 942
 943            # We need these casts to make duckdb's static type checker happy
 944            blob = exp.cast(arg, exp.DataType.Type.VARBINARY)
 945            varchar = exp.cast(arg, exp.DataType.Type.VARCHAR)
 946
 947            case = (
 948                exp.case(self.func("TYPEOF", arg))
 949                .when(
 950                    "'VARCHAR'", exp.Anonymous(this="LENGTH", expressions=[varchar])
 951                )  # anonymous to break length_sql recursion
 952                .when("'BLOB'", self.func("OCTET_LENGTH", blob))
 953            )
 954
 955            return self.sql(case)
 956
 957        def objectinsert_sql(self, expression: exp.ObjectInsert) -> str:
 958            this = expression.this
 959            key = expression.args.get("key")
 960            key_sql = key.name if isinstance(key, exp.Expression) else ""
 961            value_sql = self.sql(expression, "value")
 962
 963            kv_sql = f"{key_sql} := {value_sql}"
 964
 965            # If the input struct is empty e.g. transpiling OBJECT_INSERT(OBJECT_CONSTRUCT(), key, value) from Snowflake
 966            # then we can generate STRUCT_PACK which will build it since STRUCT_INSERT({}, key := value) is not valid DuckDB
 967            if isinstance(this, exp.Struct) and not this.expressions:
 968                return self.func("STRUCT_PACK", kv_sql)
 969
 970            return self.func("STRUCT_INSERT", this, kv_sql)
 971
 972        def unnest_sql(self, expression: exp.Unnest) -> str:
 973            explode_array = expression.args.get("explode_array")
 974            if explode_array:
 975                # In BigQuery, UNNESTing a nested array leads to explosion of the top-level array & struct
 976                # This is transpiled to DDB by transforming "FROM UNNEST(...)" to "FROM (SELECT UNNEST(..., max_depth => 2))"
 977                expression.expressions.append(
 978                    exp.Kwarg(this=exp.var("max_depth"), expression=exp.Literal.number(2))
 979                )
 980
 981                # If BQ's UNNEST is aliased, we transform it from a column alias to a table alias in DDB
 982                alias = expression.args.get("alias")
 983                if alias:
 984                    expression.set("alias", None)
 985                    alias = exp.TableAlias(this=seq_get(alias.args.get("columns"), 0))
 986
 987                unnest_sql = super().unnest_sql(expression)
 988                select = exp.Select(expressions=[unnest_sql]).subquery(alias)
 989                return self.sql(select)
 990
 991            return super().unnest_sql(expression)
 992
 993        def ignorenulls_sql(self, expression: exp.IgnoreNulls) -> str:
 994            if isinstance(expression.this, WINDOW_FUNCS_WITH_IGNORE_NULLS):
 995                # DuckDB should render IGNORE NULLS only for the general-purpose
 996                # window functions that accept it e.g. FIRST_VALUE(... IGNORE NULLS) OVER (...)
 997                return super().ignorenulls_sql(expression)
 998
 999            return self.sql(expression, "this")
1000
1001        def arraytostring_sql(self, expression: exp.ArrayToString) -> str:
1002            this = self.sql(expression, "this")
1003            null_text = self.sql(expression, "null")
1004
1005            if null_text:
1006                this = f"LIST_TRANSFORM({this}, x -> COALESCE(x, {null_text}))"
1007
1008            return self.func("ARRAY_TO_STRING", this, expression.expression)
1009
1010        @unsupported_args("position", "occurrence")
1011        def regexpextract_sql(self, expression: exp.RegexpExtract) -> str:
1012            group = expression.args.get("group")
1013            params = expression.args.get("parameters")
1014
1015            # Do not render group if there is no following argument,
1016            # and it's the default value for this dialect
1017            if (
1018                not params
1019                and group
1020                and group.name == str(self.dialect.REGEXP_EXTRACT_DEFAULT_GROUP)
1021            ):
1022                group = None
1023            return self.func(
1024                "REGEXP_EXTRACT", expression.this, expression.expression, group, params
1025            )
1026
1027        @unsupported_args("culture")
1028        def numbertostr_sql(self, expression: exp.NumberToStr) -> str:
1029            fmt = expression.args.get("format")
1030            if fmt and fmt.is_int:
1031                return self.func("FORMAT", f"'{{:,.{fmt.name}f}}'", expression.this)
1032
1033            self.unsupported("Only integer formats are supported by NumberToStr")
1034            return self.function_fallback_sql(expression)
WINDOW_FUNCS_WITH_IGNORE_NULLS = (<class 'sqlglot.expressions.FirstValue'>, <class 'sqlglot.expressions.LastValue'>, <class 'sqlglot.expressions.Lag'>, <class 'sqlglot.expressions.Lead'>, <class 'sqlglot.expressions.NthValue'>)
WRAPPED_JSON_EXTRACT_EXPRESSIONS = (<class 'sqlglot.expressions.Binary'>, <class 'sqlglot.expressions.Bracket'>, <class 'sqlglot.expressions.In'>)
class DuckDB(sqlglot.dialects.dialect.Dialect):
 274class DuckDB(Dialect):
 275    NULL_ORDERING = "nulls_are_last"
 276    SUPPORTS_USER_DEFINED_TYPES = False
 277    SAFE_DIVISION = True
 278    INDEX_OFFSET = 1
 279    CONCAT_COALESCE = True
 280    SUPPORTS_ORDER_BY_ALL = True
 281    SUPPORTS_FIXED_SIZE_ARRAYS = True
 282    STRICT_JSON_PATH_SYNTAX = False
 283
 284    # https://duckdb.org/docs/sql/introduction.html#creating-a-new-table
 285    NORMALIZATION_STRATEGY = NormalizationStrategy.CASE_INSENSITIVE
 286
 287    def to_json_path(self, path: t.Optional[exp.Expression]) -> t.Optional[exp.Expression]:
 288        if isinstance(path, exp.Literal):
 289            # DuckDB also supports the JSON pointer syntax, where every path starts with a `/`.
 290            # Additionally, it allows accessing the back of lists using the `[#-i]` syntax.
 291            # This check ensures we'll avoid trying to parse these as JSON paths, which can
 292            # either result in a noisy warning or in an invalid representation of the path.
 293            path_text = path.name
 294            if path_text.startswith("/") or "[#" in path_text:
 295                return path
 296
 297        return super().to_json_path(path)
 298
 299    class Tokenizer(tokens.Tokenizer):
 300        BYTE_STRINGS = [("e'", "'"), ("E'", "'")]
 301        HEREDOC_STRINGS = ["$"]
 302
 303        HEREDOC_TAG_IS_IDENTIFIER = True
 304        HEREDOC_STRING_ALTERNATIVE = TokenType.PARAMETER
 305
 306        KEYWORDS = {
 307            **tokens.Tokenizer.KEYWORDS,
 308            "//": TokenType.DIV,
 309            "**": TokenType.DSTAR,
 310            "^@": TokenType.CARET_AT,
 311            "@>": TokenType.AT_GT,
 312            "<@": TokenType.LT_AT,
 313            "ATTACH": TokenType.ATTACH,
 314            "BINARY": TokenType.VARBINARY,
 315            "BITSTRING": TokenType.BIT,
 316            "BPCHAR": TokenType.TEXT,
 317            "CHAR": TokenType.TEXT,
 318            "CHARACTER VARYING": TokenType.TEXT,
 319            "DETACH": TokenType.DETACH,
 320            "EXCLUDE": TokenType.EXCEPT,
 321            "LOGICAL": TokenType.BOOLEAN,
 322            "ONLY": TokenType.ONLY,
 323            "PIVOT_WIDER": TokenType.PIVOT,
 324            "POSITIONAL": TokenType.POSITIONAL,
 325            "SIGNED": TokenType.INT,
 326            "STRING": TokenType.TEXT,
 327            "SUMMARIZE": TokenType.SUMMARIZE,
 328            "TIMESTAMP_S": TokenType.TIMESTAMP_S,
 329            "TIMESTAMP_MS": TokenType.TIMESTAMP_MS,
 330            "TIMESTAMP_NS": TokenType.TIMESTAMP_NS,
 331            "TIMESTAMP_US": TokenType.TIMESTAMP,
 332            "UBIGINT": TokenType.UBIGINT,
 333            "UINTEGER": TokenType.UINT,
 334            "USMALLINT": TokenType.USMALLINT,
 335            "UTINYINT": TokenType.UTINYINT,
 336            "VARCHAR": TokenType.TEXT,
 337        }
 338        KEYWORDS.pop("/*+")
 339
 340        SINGLE_TOKENS = {
 341            **tokens.Tokenizer.SINGLE_TOKENS,
 342            "$": TokenType.PARAMETER,
 343        }
 344
 345    class Parser(parser.Parser):
 346        BITWISE = {
 347            **parser.Parser.BITWISE,
 348            TokenType.TILDA: exp.RegexpLike,
 349        }
 350        BITWISE.pop(TokenType.CARET)
 351
 352        RANGE_PARSERS = {
 353            **parser.Parser.RANGE_PARSERS,
 354            TokenType.DAMP: binary_range_parser(exp.ArrayOverlaps),
 355            TokenType.CARET_AT: binary_range_parser(exp.StartsWith),
 356        }
 357
 358        EXPONENT = {
 359            **parser.Parser.EXPONENT,
 360            TokenType.CARET: exp.Pow,
 361            TokenType.DSTAR: exp.Pow,
 362        }
 363
 364        FUNCTIONS_WITH_ALIASED_ARGS = {*parser.Parser.FUNCTIONS_WITH_ALIASED_ARGS, "STRUCT_PACK"}
 365
 366        FUNCTIONS = {
 367            **parser.Parser.FUNCTIONS,
 368            "ARRAY_REVERSE_SORT": _build_sort_array_desc,
 369            "ARRAY_SORT": exp.SortArray.from_arg_list,
 370            "DATEDIFF": _build_date_diff,
 371            "DATE_DIFF": _build_date_diff,
 372            "DATE_TRUNC": date_trunc_to_time,
 373            "DATETRUNC": date_trunc_to_time,
 374            "DECODE": lambda args: exp.Decode(
 375                this=seq_get(args, 0), charset=exp.Literal.string("utf-8")
 376            ),
 377            "ENCODE": lambda args: exp.Encode(
 378                this=seq_get(args, 0), charset=exp.Literal.string("utf-8")
 379            ),
 380            "EPOCH": exp.TimeToUnix.from_arg_list,
 381            "EPOCH_MS": lambda args: exp.UnixToTime(
 382                this=seq_get(args, 0), scale=exp.UnixToTime.MILLIS
 383            ),
 384            "JSON": exp.ParseJSON.from_arg_list,
 385            "JSON_EXTRACT_PATH": parser.build_extract_json_with_path(exp.JSONExtract),
 386            "JSON_EXTRACT_STRING": parser.build_extract_json_with_path(exp.JSONExtractScalar),
 387            "LIST_HAS": exp.ArrayContains.from_arg_list,
 388            "LIST_REVERSE_SORT": _build_sort_array_desc,
 389            "LIST_SORT": exp.SortArray.from_arg_list,
 390            "LIST_VALUE": lambda args: exp.Array(expressions=args),
 391            "MAKE_TIME": exp.TimeFromParts.from_arg_list,
 392            "MAKE_TIMESTAMP": _build_make_timestamp,
 393            "QUANTILE_CONT": exp.PercentileCont.from_arg_list,
 394            "QUANTILE_DISC": exp.PercentileDisc.from_arg_list,
 395            "REGEXP_EXTRACT": build_regexp_extract(exp.RegexpExtract),
 396            "REGEXP_EXTRACT_ALL": build_regexp_extract(exp.RegexpExtractAll),
 397            "REGEXP_MATCHES": exp.RegexpLike.from_arg_list,
 398            "REGEXP_REPLACE": lambda args: exp.RegexpReplace(
 399                this=seq_get(args, 0),
 400                expression=seq_get(args, 1),
 401                replacement=seq_get(args, 2),
 402                modifiers=seq_get(args, 3),
 403            ),
 404            "STRFTIME": build_formatted_time(exp.TimeToStr, "duckdb"),
 405            "STRING_SPLIT": exp.Split.from_arg_list,
 406            "STRING_SPLIT_REGEX": exp.RegexpSplit.from_arg_list,
 407            "STRING_TO_ARRAY": exp.Split.from_arg_list,
 408            "STRPTIME": build_formatted_time(exp.StrToTime, "duckdb"),
 409            "STRUCT_PACK": exp.Struct.from_arg_list,
 410            "STR_SPLIT": exp.Split.from_arg_list,
 411            "STR_SPLIT_REGEX": exp.RegexpSplit.from_arg_list,
 412            "TO_TIMESTAMP": exp.UnixToTime.from_arg_list,
 413            "UNNEST": exp.Explode.from_arg_list,
 414            "XOR": binary_from_function(exp.BitwiseXor),
 415            "GENERATE_SERIES": _build_generate_series(),
 416            "RANGE": _build_generate_series(end_exclusive=True),
 417            "EDITDIST3": exp.Levenshtein.from_arg_list,
 418        }
 419
 420        FUNCTIONS.pop("DATE_SUB")
 421        FUNCTIONS.pop("GLOB")
 422
 423        FUNCTION_PARSERS = parser.Parser.FUNCTION_PARSERS.copy()
 424        FUNCTION_PARSERS.pop("DECODE")
 425
 426        NO_PAREN_FUNCTION_PARSERS = {
 427            **parser.Parser.NO_PAREN_FUNCTION_PARSERS,
 428            "MAP": lambda self: self._parse_map(),
 429        }
 430
 431        TABLE_ALIAS_TOKENS = parser.Parser.TABLE_ALIAS_TOKENS - {
 432            TokenType.SEMI,
 433            TokenType.ANTI,
 434        }
 435
 436        PLACEHOLDER_PARSERS = {
 437            **parser.Parser.PLACEHOLDER_PARSERS,
 438            TokenType.PARAMETER: lambda self: (
 439                self.expression(exp.Placeholder, this=self._prev.text)
 440                if self._match(TokenType.NUMBER) or self._match_set(self.ID_VAR_TOKENS)
 441                else None
 442            ),
 443        }
 444
 445        TYPE_CONVERTERS = {
 446            # https://duckdb.org/docs/sql/data_types/numeric
 447            exp.DataType.Type.DECIMAL: build_default_decimal_type(precision=18, scale=3),
 448            # https://duckdb.org/docs/sql/data_types/text
 449            exp.DataType.Type.TEXT: lambda dtype: exp.DataType.build("TEXT"),
 450        }
 451
 452        STATEMENT_PARSERS = {
 453            **parser.Parser.STATEMENT_PARSERS,
 454            TokenType.ATTACH: lambda self: self._parse_attach_detach(),
 455            TokenType.DETACH: lambda self: self._parse_attach_detach(is_attach=False),
 456        }
 457
 458        def _parse_table_sample(self, as_modifier: bool = False) -> t.Optional[exp.TableSample]:
 459            # https://duckdb.org/docs/sql/samples.html
 460            sample = super()._parse_table_sample(as_modifier=as_modifier)
 461            if sample and not sample.args.get("method"):
 462                if sample.args.get("size"):
 463                    sample.set("method", exp.var("RESERVOIR"))
 464                else:
 465                    sample.set("method", exp.var("SYSTEM"))
 466
 467            return sample
 468
 469        def _parse_bracket(
 470            self, this: t.Optional[exp.Expression] = None
 471        ) -> t.Optional[exp.Expression]:
 472            bracket = super()._parse_bracket(this)
 473            if isinstance(bracket, exp.Bracket):
 474                bracket.set("returns_list_for_maps", True)
 475
 476            return bracket
 477
 478        def _parse_map(self) -> exp.ToMap | exp.Map:
 479            if self._match(TokenType.L_BRACE, advance=False):
 480                return self.expression(exp.ToMap, this=self._parse_bracket())
 481
 482            args = self._parse_wrapped_csv(self._parse_assignment)
 483            return self.expression(exp.Map, keys=seq_get(args, 0), values=seq_get(args, 1))
 484
 485        def _parse_struct_types(self, type_required: bool = False) -> t.Optional[exp.Expression]:
 486            return self._parse_field_def()
 487
 488        def _pivot_column_names(self, aggregations: t.List[exp.Expression]) -> t.List[str]:
 489            if len(aggregations) == 1:
 490                return super()._pivot_column_names(aggregations)
 491            return pivot_column_names(aggregations, dialect="duckdb")
 492
 493        def _parse_attach_detach(self, is_attach=True) -> exp.Attach | exp.Detach:
 494            def _parse_attach_option() -> exp.AttachOption:
 495                return self.expression(
 496                    exp.AttachOption,
 497                    this=self._parse_var(any_token=True),
 498                    expression=self._parse_field(any_token=True),
 499                )
 500
 501            self._match(TokenType.DATABASE)
 502            exists = self._parse_exists(not_=is_attach)
 503            this = self._parse_alias(self._parse_primary_or_var(), explicit=True)
 504
 505            if self._match(TokenType.L_PAREN, advance=False):
 506                expressions = self._parse_wrapped_csv(_parse_attach_option)
 507            else:
 508                expressions = None
 509
 510            return (
 511                self.expression(exp.Attach, this=this, exists=exists, expressions=expressions)
 512                if is_attach
 513                else self.expression(exp.Detach, this=this, exists=exists)
 514            )
 515
 516    class Generator(generator.Generator):
 517        PARAMETER_TOKEN = "$"
 518        NAMED_PLACEHOLDER_TOKEN = "$"
 519        JOIN_HINTS = False
 520        TABLE_HINTS = False
 521        QUERY_HINTS = False
 522        LIMIT_FETCH = "LIMIT"
 523        STRUCT_DELIMITER = ("(", ")")
 524        RENAME_TABLE_WITH_DB = False
 525        NVL2_SUPPORTED = False
 526        SEMI_ANTI_JOIN_WITH_SIDE = False
 527        TABLESAMPLE_KEYWORDS = "USING SAMPLE"
 528        TABLESAMPLE_SEED_KEYWORD = "REPEATABLE"
 529        LAST_DAY_SUPPORTS_DATE_PART = False
 530        JSON_KEY_VALUE_PAIR_SEP = ","
 531        IGNORE_NULLS_IN_FUNC = True
 532        JSON_PATH_BRACKETED_KEY_SUPPORTED = False
 533        SUPPORTS_CREATE_TABLE_LIKE = False
 534        MULTI_ARG_DISTINCT = False
 535        CAN_IMPLEMENT_ARRAY_ANY = True
 536        SUPPORTS_TO_NUMBER = False
 537        COPY_HAS_INTO_KEYWORD = False
 538        STAR_EXCEPT = "EXCLUDE"
 539        PAD_FILL_PATTERN_IS_REQUIRED = True
 540        ARRAY_CONCAT_IS_VAR_LEN = False
 541        ARRAY_SIZE_DIM_REQUIRED = False
 542
 543        TRANSFORMS = {
 544            **generator.Generator.TRANSFORMS,
 545            exp.ApproxDistinct: approx_count_distinct_sql,
 546            exp.Array: inline_array_unless_query,
 547            exp.ArrayFilter: rename_func("LIST_FILTER"),
 548            exp.ArraySort: _array_sort_sql,
 549            exp.ArraySum: rename_func("LIST_SUM"),
 550            exp.BitwiseXor: rename_func("XOR"),
 551            exp.CommentColumnConstraint: no_comment_column_constraint_sql,
 552            exp.CurrentDate: lambda *_: "CURRENT_DATE",
 553            exp.CurrentTime: lambda *_: "CURRENT_TIME",
 554            exp.CurrentTimestamp: lambda *_: "CURRENT_TIMESTAMP",
 555            exp.DayOfMonth: rename_func("DAYOFMONTH"),
 556            exp.DayOfWeek: rename_func("DAYOFWEEK"),
 557            exp.DayOfWeekIso: rename_func("ISODOW"),
 558            exp.DayOfYear: rename_func("DAYOFYEAR"),
 559            exp.DataType: _datatype_sql,
 560            exp.Date: _date_sql,
 561            exp.DateAdd: _date_delta_sql,
 562            exp.DateFromParts: rename_func("MAKE_DATE"),
 563            exp.DateSub: _date_delta_sql,
 564            exp.DateDiff: _date_diff_sql,
 565            exp.DateStrToDate: datestrtodate_sql,
 566            exp.Datetime: no_datetime_sql,
 567            exp.DatetimeSub: _date_delta_sql,
 568            exp.DatetimeAdd: _date_delta_sql,
 569            exp.DateToDi: lambda self,
 570            e: f"CAST(STRFTIME({self.sql(e, 'this')}, {DuckDB.DATEINT_FORMAT}) AS INT)",
 571            exp.Decode: lambda self, e: encode_decode_sql(self, e, "DECODE", replace=False),
 572            exp.DiToDate: lambda self,
 573            e: f"CAST(STRPTIME(CAST({self.sql(e, 'this')} AS TEXT), {DuckDB.DATEINT_FORMAT}) AS DATE)",
 574            exp.Encode: lambda self, e: encode_decode_sql(self, e, "ENCODE", replace=False),
 575            exp.GenerateDateArray: _generate_datetime_array_sql,
 576            exp.GenerateTimestampArray: _generate_datetime_array_sql,
 577            exp.Explode: rename_func("UNNEST"),
 578            exp.IntDiv: lambda self, e: self.binary(e, "//"),
 579            exp.IsInf: rename_func("ISINF"),
 580            exp.IsNan: rename_func("ISNAN"),
 581            exp.JSONBExists: rename_func("JSON_EXISTS"),
 582            exp.JSONExtract: _arrow_json_extract_sql,
 583            exp.JSONExtractArray: _json_extract_value_array_sql,
 584            exp.JSONExtractScalar: _arrow_json_extract_sql,
 585            exp.JSONFormat: _json_format_sql,
 586            exp.JSONValueArray: _json_extract_value_array_sql,
 587            exp.Lateral: explode_to_unnest_sql,
 588            exp.LogicalOr: rename_func("BOOL_OR"),
 589            exp.LogicalAnd: rename_func("BOOL_AND"),
 590            exp.MakeInterval: lambda self, e: no_make_interval_sql(self, e, sep=" "),
 591            exp.MD5Digest: lambda self, e: self.func("UNHEX", self.func("MD5", e.this)),
 592            exp.MonthsBetween: lambda self, e: self.func(
 593                "DATEDIFF",
 594                "'month'",
 595                exp.cast(e.expression, exp.DataType.Type.TIMESTAMP, copy=True),
 596                exp.cast(e.this, exp.DataType.Type.TIMESTAMP, copy=True),
 597            ),
 598            exp.PercentileCont: rename_func("QUANTILE_CONT"),
 599            exp.PercentileDisc: rename_func("QUANTILE_DISC"),
 600            # DuckDB doesn't allow qualified columns inside of PIVOT expressions.
 601            # See: https://github.com/duckdb/duckdb/blob/671faf92411182f81dce42ac43de8bfb05d9909e/src/planner/binder/tableref/bind_pivot.cpp#L61-L62
 602            exp.Pivot: transforms.preprocess([transforms.unqualify_columns]),
 603            exp.RegexpReplace: lambda self, e: self.func(
 604                "REGEXP_REPLACE",
 605                e.this,
 606                e.expression,
 607                e.args.get("replacement"),
 608                e.args.get("modifiers"),
 609            ),
 610            exp.RegexpLike: rename_func("REGEXP_MATCHES"),
 611            exp.RegexpILike: lambda self, e: self.func(
 612                "REGEXP_MATCHES", e.this, e.expression, exp.Literal.string("i")
 613            ),
 614            exp.RegexpSplit: rename_func("STR_SPLIT_REGEX"),
 615            exp.Return: lambda self, e: self.sql(e, "this"),
 616            exp.ReturnsProperty: lambda self, e: "TABLE" if isinstance(e.this, exp.Schema) else "",
 617            exp.Rand: rename_func("RANDOM"),
 618            exp.SHA: rename_func("SHA1"),
 619            exp.SHA2: sha256_sql,
 620            exp.Split: rename_func("STR_SPLIT"),
 621            exp.SortArray: _sort_array_sql,
 622            exp.StrPosition: strposition_sql,
 623            exp.StrToUnix: lambda self, e: self.func(
 624                "EPOCH", self.func("STRPTIME", e.this, self.format_time(e))
 625            ),
 626            exp.Struct: _struct_sql,
 627            exp.Transform: rename_func("LIST_TRANSFORM"),
 628            exp.TimeAdd: _date_delta_sql,
 629            exp.Time: no_time_sql,
 630            exp.TimeDiff: _timediff_sql,
 631            exp.Timestamp: no_timestamp_sql,
 632            exp.TimestampDiff: lambda self, e: self.func(
 633                "DATE_DIFF", exp.Literal.string(e.unit), e.expression, e.this
 634            ),
 635            exp.TimestampTrunc: timestamptrunc_sql(),
 636            exp.TimeStrToDate: lambda self, e: self.sql(exp.cast(e.this, exp.DataType.Type.DATE)),
 637            exp.TimeStrToTime: timestrtotime_sql,
 638            exp.TimeStrToUnix: lambda self, e: self.func(
 639                "EPOCH", exp.cast(e.this, exp.DataType.Type.TIMESTAMP)
 640            ),
 641            exp.TimeToStr: lambda self, e: self.func("STRFTIME", e.this, self.format_time(e)),
 642            exp.TimeToUnix: rename_func("EPOCH"),
 643            exp.TsOrDiToDi: lambda self,
 644            e: f"CAST(SUBSTR(REPLACE(CAST({self.sql(e, 'this')} AS TEXT), '-', ''), 1, 8) AS INT)",
 645            exp.TsOrDsAdd: _date_delta_sql,
 646            exp.TsOrDsDiff: lambda self, e: self.func(
 647                "DATE_DIFF",
 648                f"'{e.args.get('unit') or 'DAY'}'",
 649                exp.cast(e.expression, exp.DataType.Type.TIMESTAMP),
 650                exp.cast(e.this, exp.DataType.Type.TIMESTAMP),
 651            ),
 652            exp.UnixToStr: lambda self, e: self.func(
 653                "STRFTIME", self.func("TO_TIMESTAMP", e.this), self.format_time(e)
 654            ),
 655            exp.DatetimeTrunc: lambda self, e: self.func(
 656                "DATE_TRUNC", unit_to_str(e), exp.cast(e.this, exp.DataType.Type.DATETIME)
 657            ),
 658            exp.UnixToTime: _unix_to_time_sql,
 659            exp.UnixToTimeStr: lambda self, e: f"CAST(TO_TIMESTAMP({self.sql(e, 'this')}) AS TEXT)",
 660            exp.VariancePop: rename_func("VAR_POP"),
 661            exp.WeekOfYear: rename_func("WEEKOFYEAR"),
 662            exp.Xor: bool_xor_sql,
 663            exp.Levenshtein: unsupported_args("ins_cost", "del_cost", "sub_cost", "max_dist")(
 664                rename_func("LEVENSHTEIN")
 665            ),
 666        }
 667
 668        SUPPORTED_JSON_PATH_PARTS = {
 669            exp.JSONPathKey,
 670            exp.JSONPathRoot,
 671            exp.JSONPathSubscript,
 672            exp.JSONPathWildcard,
 673        }
 674
 675        TYPE_MAPPING = {
 676            **generator.Generator.TYPE_MAPPING,
 677            exp.DataType.Type.BINARY: "BLOB",
 678            exp.DataType.Type.BPCHAR: "TEXT",
 679            exp.DataType.Type.CHAR: "TEXT",
 680            exp.DataType.Type.DATETIME: "TIMESTAMP",
 681            exp.DataType.Type.FLOAT: "REAL",
 682            exp.DataType.Type.NCHAR: "TEXT",
 683            exp.DataType.Type.NVARCHAR: "TEXT",
 684            exp.DataType.Type.UINT: "UINTEGER",
 685            exp.DataType.Type.VARBINARY: "BLOB",
 686            exp.DataType.Type.ROWVERSION: "BLOB",
 687            exp.DataType.Type.VARCHAR: "TEXT",
 688            exp.DataType.Type.TIMESTAMPNTZ: "TIMESTAMP",
 689            exp.DataType.Type.TIMESTAMP_S: "TIMESTAMP_S",
 690            exp.DataType.Type.TIMESTAMP_MS: "TIMESTAMP_MS",
 691            exp.DataType.Type.TIMESTAMP_NS: "TIMESTAMP_NS",
 692        }
 693
 694        # https://github.com/duckdb/duckdb/blob/ff7f24fd8e3128d94371827523dae85ebaf58713/third_party/libpg_query/grammar/keywords/reserved_keywords.list#L1-L77
 695        RESERVED_KEYWORDS = {
 696            "array",
 697            "analyse",
 698            "union",
 699            "all",
 700            "when",
 701            "in_p",
 702            "default",
 703            "create_p",
 704            "window",
 705            "asymmetric",
 706            "to",
 707            "else",
 708            "localtime",
 709            "from",
 710            "end_p",
 711            "select",
 712            "current_date",
 713            "foreign",
 714            "with",
 715            "grant",
 716            "session_user",
 717            "or",
 718            "except",
 719            "references",
 720            "fetch",
 721            "limit",
 722            "group_p",
 723            "leading",
 724            "into",
 725            "collate",
 726            "offset",
 727            "do",
 728            "then",
 729            "localtimestamp",
 730            "check_p",
 731            "lateral_p",
 732            "current_role",
 733            "where",
 734            "asc_p",
 735            "placing",
 736            "desc_p",
 737            "user",
 738            "unique",
 739            "initially",
 740            "column",
 741            "both",
 742            "some",
 743            "as",
 744            "any",
 745            "only",
 746            "deferrable",
 747            "null_p",
 748            "current_time",
 749            "true_p",
 750            "table",
 751            "case",
 752            "trailing",
 753            "variadic",
 754            "for",
 755            "on",
 756            "distinct",
 757            "false_p",
 758            "not",
 759            "constraint",
 760            "current_timestamp",
 761            "returning",
 762            "primary",
 763            "intersect",
 764            "having",
 765            "analyze",
 766            "current_user",
 767            "and",
 768            "cast",
 769            "symmetric",
 770            "using",
 771            "order",
 772            "current_catalog",
 773        }
 774
 775        UNWRAPPED_INTERVAL_VALUES = (exp.Literal, exp.Paren)
 776
 777        # DuckDB doesn't generally support CREATE TABLE .. properties
 778        # https://duckdb.org/docs/sql/statements/create_table.html
 779        PROPERTIES_LOCATION = {
 780            prop: exp.Properties.Location.UNSUPPORTED
 781            for prop in generator.Generator.PROPERTIES_LOCATION
 782        }
 783
 784        # There are a few exceptions (e.g. temporary tables) which are supported or
 785        # can be transpiled to DuckDB, so we explicitly override them accordingly
 786        PROPERTIES_LOCATION[exp.LikeProperty] = exp.Properties.Location.POST_SCHEMA
 787        PROPERTIES_LOCATION[exp.TemporaryProperty] = exp.Properties.Location.POST_CREATE
 788        PROPERTIES_LOCATION[exp.ReturnsProperty] = exp.Properties.Location.POST_ALIAS
 789
 790        def fromiso8601timestamp_sql(self, expression: exp.FromISO8601Timestamp) -> str:
 791            return self.sql(exp.cast(expression.this, exp.DataType.Type.TIMESTAMPTZ))
 792
 793        def strtotime_sql(self, expression: exp.StrToTime) -> str:
 794            if expression.args.get("safe"):
 795                formatted_time = self.format_time(expression)
 796                return f"CAST({self.func('TRY_STRPTIME', expression.this, formatted_time)} AS TIMESTAMP)"
 797            return str_to_time_sql(self, expression)
 798
 799        def strtodate_sql(self, expression: exp.StrToDate) -> str:
 800            if expression.args.get("safe"):
 801                formatted_time = self.format_time(expression)
 802                return f"CAST({self.func('TRY_STRPTIME', expression.this, formatted_time)} AS DATE)"
 803            return f"CAST({str_to_time_sql(self, expression)} AS DATE)"
 804
 805        def parsejson_sql(self, expression: exp.ParseJSON) -> str:
 806            arg = expression.this
 807            if expression.args.get("safe"):
 808                return self.sql(exp.case().when(exp.func("json_valid", arg), arg).else_(exp.null()))
 809            return self.func("JSON", arg)
 810
 811        def timefromparts_sql(self, expression: exp.TimeFromParts) -> str:
 812            nano = expression.args.get("nano")
 813            if nano is not None:
 814                expression.set(
 815                    "sec", expression.args["sec"] + nano.pop() / exp.Literal.number(1000000000.0)
 816                )
 817
 818            return rename_func("MAKE_TIME")(self, expression)
 819
 820        def timestampfromparts_sql(self, expression: exp.TimestampFromParts) -> str:
 821            sec = expression.args["sec"]
 822
 823            milli = expression.args.get("milli")
 824            if milli is not None:
 825                sec += milli.pop() / exp.Literal.number(1000.0)
 826
 827            nano = expression.args.get("nano")
 828            if nano is not None:
 829                sec += nano.pop() / exp.Literal.number(1000000000.0)
 830
 831            if milli or nano:
 832                expression.set("sec", sec)
 833
 834            return rename_func("MAKE_TIMESTAMP")(self, expression)
 835
 836        def tablesample_sql(
 837            self,
 838            expression: exp.TableSample,
 839            tablesample_keyword: t.Optional[str] = None,
 840        ) -> str:
 841            if not isinstance(expression.parent, exp.Select):
 842                # This sample clause only applies to a single source, not the entire resulting relation
 843                tablesample_keyword = "TABLESAMPLE"
 844
 845            if expression.args.get("size"):
 846                method = expression.args.get("method")
 847                if method and method.name.upper() != "RESERVOIR":
 848                    self.unsupported(
 849                        f"Sampling method {method} is not supported with a discrete sample count, "
 850                        "defaulting to reservoir sampling"
 851                    )
 852                    expression.set("method", exp.var("RESERVOIR"))
 853
 854            return super().tablesample_sql(expression, tablesample_keyword=tablesample_keyword)
 855
 856        def interval_sql(self, expression: exp.Interval) -> str:
 857            multiplier: t.Optional[int] = None
 858            unit = expression.text("unit").lower()
 859
 860            if unit.startswith("week"):
 861                multiplier = 7
 862            if unit.startswith("quarter"):
 863                multiplier = 90
 864
 865            if multiplier:
 866                return f"({multiplier} * {super().interval_sql(exp.Interval(this=expression.this, unit=exp.var('DAY')))})"
 867
 868            return super().interval_sql(expression)
 869
 870        def columndef_sql(self, expression: exp.ColumnDef, sep: str = " ") -> str:
 871            if isinstance(expression.parent, exp.UserDefinedFunction):
 872                return self.sql(expression, "this")
 873            return super().columndef_sql(expression, sep)
 874
 875        def join_sql(self, expression: exp.Join) -> str:
 876            if (
 877                expression.side == "LEFT"
 878                and not expression.args.get("on")
 879                and isinstance(expression.this, exp.Unnest)
 880            ):
 881                # Some dialects support `LEFT JOIN UNNEST(...)` without an explicit ON clause
 882                # DuckDB doesn't, but we can just add a dummy ON clause that is always true
 883                return super().join_sql(expression.on(exp.true()))
 884
 885            return super().join_sql(expression)
 886
 887        def generateseries_sql(self, expression: exp.GenerateSeries) -> str:
 888            # GENERATE_SERIES(a, b) -> [a, b], RANGE(a, b) -> [a, b)
 889            if expression.args.get("is_end_exclusive"):
 890                return rename_func("RANGE")(self, expression)
 891
 892            return self.function_fallback_sql(expression)
 893
 894        def bracket_sql(self, expression: exp.Bracket) -> str:
 895            this = expression.this
 896            if isinstance(this, exp.Array):
 897                this.replace(exp.paren(this))
 898
 899            bracket = super().bracket_sql(expression)
 900
 901            if not expression.args.get("returns_list_for_maps"):
 902                if not this.type:
 903                    from sqlglot.optimizer.annotate_types import annotate_types
 904
 905                    this = annotate_types(this)
 906
 907                if this.is_type(exp.DataType.Type.MAP):
 908                    bracket = f"({bracket})[1]"
 909
 910            return bracket
 911
 912        def withingroup_sql(self, expression: exp.WithinGroup) -> str:
 913            expression_sql = self.sql(expression, "expression")
 914
 915            func = expression.this
 916            if isinstance(func, exp.PERCENTILES):
 917                # Make the order key the first arg and slide the fraction to the right
 918                # https://duckdb.org/docs/sql/aggregates#ordered-set-aggregate-functions
 919                order_col = expression.find(exp.Ordered)
 920                if order_col:
 921                    func.set("expression", func.this)
 922                    func.set("this", order_col.this)
 923
 924            this = self.sql(expression, "this").rstrip(")")
 925
 926            return f"{this}{expression_sql})"
 927
 928        def length_sql(self, expression: exp.Length) -> str:
 929            arg = expression.this
 930
 931            # Dialects like BQ and Snowflake also accept binary values as args, so
 932            # DDB will attempt to infer the type or resort to case/when resolution
 933            if not expression.args.get("binary") or arg.is_string:
 934                return self.func("LENGTH", arg)
 935
 936            if not arg.type:
 937                from sqlglot.optimizer.annotate_types import annotate_types
 938
 939                arg = annotate_types(arg)
 940
 941            if arg.is_type(*exp.DataType.TEXT_TYPES):
 942                return self.func("LENGTH", arg)
 943
 944            # We need these casts to make duckdb's static type checker happy
 945            blob = exp.cast(arg, exp.DataType.Type.VARBINARY)
 946            varchar = exp.cast(arg, exp.DataType.Type.VARCHAR)
 947
 948            case = (
 949                exp.case(self.func("TYPEOF", arg))
 950                .when(
 951                    "'VARCHAR'", exp.Anonymous(this="LENGTH", expressions=[varchar])
 952                )  # anonymous to break length_sql recursion
 953                .when("'BLOB'", self.func("OCTET_LENGTH", blob))
 954            )
 955
 956            return self.sql(case)
 957
 958        def objectinsert_sql(self, expression: exp.ObjectInsert) -> str:
 959            this = expression.this
 960            key = expression.args.get("key")
 961            key_sql = key.name if isinstance(key, exp.Expression) else ""
 962            value_sql = self.sql(expression, "value")
 963
 964            kv_sql = f"{key_sql} := {value_sql}"
 965
 966            # If the input struct is empty e.g. transpiling OBJECT_INSERT(OBJECT_CONSTRUCT(), key, value) from Snowflake
 967            # then we can generate STRUCT_PACK which will build it since STRUCT_INSERT({}, key := value) is not valid DuckDB
 968            if isinstance(this, exp.Struct) and not this.expressions:
 969                return self.func("STRUCT_PACK", kv_sql)
 970
 971            return self.func("STRUCT_INSERT", this, kv_sql)
 972
 973        def unnest_sql(self, expression: exp.Unnest) -> str:
 974            explode_array = expression.args.get("explode_array")
 975            if explode_array:
 976                # In BigQuery, UNNESTing a nested array leads to explosion of the top-level array & struct
 977                # This is transpiled to DDB by transforming "FROM UNNEST(...)" to "FROM (SELECT UNNEST(..., max_depth => 2))"
 978                expression.expressions.append(
 979                    exp.Kwarg(this=exp.var("max_depth"), expression=exp.Literal.number(2))
 980                )
 981
 982                # If BQ's UNNEST is aliased, we transform it from a column alias to a table alias in DDB
 983                alias = expression.args.get("alias")
 984                if alias:
 985                    expression.set("alias", None)
 986                    alias = exp.TableAlias(this=seq_get(alias.args.get("columns"), 0))
 987
 988                unnest_sql = super().unnest_sql(expression)
 989                select = exp.Select(expressions=[unnest_sql]).subquery(alias)
 990                return self.sql(select)
 991
 992            return super().unnest_sql(expression)
 993
 994        def ignorenulls_sql(self, expression: exp.IgnoreNulls) -> str:
 995            if isinstance(expression.this, WINDOW_FUNCS_WITH_IGNORE_NULLS):
 996                # DuckDB should render IGNORE NULLS only for the general-purpose
 997                # window functions that accept it e.g. FIRST_VALUE(... IGNORE NULLS) OVER (...)
 998                return super().ignorenulls_sql(expression)
 999
1000            return self.sql(expression, "this")
1001
1002        def arraytostring_sql(self, expression: exp.ArrayToString) -> str:
1003            this = self.sql(expression, "this")
1004            null_text = self.sql(expression, "null")
1005
1006            if null_text:
1007                this = f"LIST_TRANSFORM({this}, x -> COALESCE(x, {null_text}))"
1008
1009            return self.func("ARRAY_TO_STRING", this, expression.expression)
1010
1011        @unsupported_args("position", "occurrence")
1012        def regexpextract_sql(self, expression: exp.RegexpExtract) -> str:
1013            group = expression.args.get("group")
1014            params = expression.args.get("parameters")
1015
1016            # Do not render group if there is no following argument,
1017            # and it's the default value for this dialect
1018            if (
1019                not params
1020                and group
1021                and group.name == str(self.dialect.REGEXP_EXTRACT_DEFAULT_GROUP)
1022            ):
1023                group = None
1024            return self.func(
1025                "REGEXP_EXTRACT", expression.this, expression.expression, group, params
1026            )
1027
1028        @unsupported_args("culture")
1029        def numbertostr_sql(self, expression: exp.NumberToStr) -> str:
1030            fmt = expression.args.get("format")
1031            if fmt and fmt.is_int:
1032                return self.func("FORMAT", f"'{{:,.{fmt.name}f}}'", expression.this)
1033
1034            self.unsupported("Only integer formats are supported by NumberToStr")
1035            return self.function_fallback_sql(expression)
NULL_ORDERING = 'nulls_are_last'

Default NULL ordering method to use if not explicitly set. Possible values: "nulls_are_small", "nulls_are_large", "nulls_are_last"

SUPPORTS_USER_DEFINED_TYPES = False

Whether user-defined data types are supported.

SAFE_DIVISION = True

Whether division by zero throws an error (False) or returns NULL (True).

INDEX_OFFSET = 1

The base index offset for arrays.

CONCAT_COALESCE = True

A NULL arg in CONCAT yields NULL by default, but in some dialects it yields an empty string.

SUPPORTS_ORDER_BY_ALL = True

Whether ORDER BY ALL is supported (expands to all the selected columns) as in DuckDB, Spark3/Databricks

SUPPORTS_FIXED_SIZE_ARRAYS = True

Whether expressions such as x::INT[5] should be parsed as fixed-size array defs/casts e.g. in DuckDB. In dialects which don't support fixed size arrays such as Snowflake, this should be interpreted as a subscript/index operator.

STRICT_JSON_PATH_SYNTAX = False

Whether failing to parse a JSON path expression using the JSONPath dialect will log a warning.

NORMALIZATION_STRATEGY = <NormalizationStrategy.CASE_INSENSITIVE: 'CASE_INSENSITIVE'>

Specifies the strategy according to which identifiers should be normalized.

def to_json_path( self, path: Optional[sqlglot.expressions.Expression]) -> Optional[sqlglot.expressions.Expression]:
287    def to_json_path(self, path: t.Optional[exp.Expression]) -> t.Optional[exp.Expression]:
288        if isinstance(path, exp.Literal):
289            # DuckDB also supports the JSON pointer syntax, where every path starts with a `/`.
290            # Additionally, it allows accessing the back of lists using the `[#-i]` syntax.
291            # This check ensures we'll avoid trying to parse these as JSON paths, which can
292            # either result in a noisy warning or in an invalid representation of the path.
293            path_text = path.name
294            if path_text.startswith("/") or "[#" in path_text:
295                return path
296
297        return super().to_json_path(path)
SUPPORTS_COLUMN_JOIN_MARKS = False

Whether the old-style outer join (+) syntax is supported.

tokenizer_class = <class 'DuckDB.Tokenizer'>
jsonpath_tokenizer_class = <class 'sqlglot.tokens.JSONPathTokenizer'>
parser_class = <class 'DuckDB.Parser'>
generator_class = <class 'DuckDB.Generator'>
TIME_TRIE: Dict = {}
FORMAT_TRIE: Dict = {}
INVERSE_TIME_MAPPING: Dict[str, str] = {}
INVERSE_TIME_TRIE: Dict = {}
INVERSE_FORMAT_MAPPING: Dict[str, str] = {}
INVERSE_FORMAT_TRIE: Dict = {}
INVERSE_CREATABLE_KIND_MAPPING: dict[str, str] = {}
ESCAPED_SEQUENCES: Dict[str, str] = {}
QUOTE_START = "'"
QUOTE_END = "'"
IDENTIFIER_START = '"'
IDENTIFIER_END = '"'
BIT_START: Optional[str] = None
BIT_END: Optional[str] = None
HEX_START: Optional[str] = None
HEX_END: Optional[str] = None
BYTE_START: Optional[str] = "e'"
BYTE_END: Optional[str] = "'"
UNICODE_START: Optional[str] = None
UNICODE_END: Optional[str] = None
class DuckDB.Tokenizer(sqlglot.tokens.Tokenizer):
299    class Tokenizer(tokens.Tokenizer):
300        BYTE_STRINGS = [("e'", "'"), ("E'", "'")]
301        HEREDOC_STRINGS = ["$"]
302
303        HEREDOC_TAG_IS_IDENTIFIER = True
304        HEREDOC_STRING_ALTERNATIVE = TokenType.PARAMETER
305
306        KEYWORDS = {
307            **tokens.Tokenizer.KEYWORDS,
308            "//": TokenType.DIV,
309            "**": TokenType.DSTAR,
310            "^@": TokenType.CARET_AT,
311            "@>": TokenType.AT_GT,
312            "<@": TokenType.LT_AT,
313            "ATTACH": TokenType.ATTACH,
314            "BINARY": TokenType.VARBINARY,
315            "BITSTRING": TokenType.BIT,
316            "BPCHAR": TokenType.TEXT,
317            "CHAR": TokenType.TEXT,
318            "CHARACTER VARYING": TokenType.TEXT,
319            "DETACH": TokenType.DETACH,
320            "EXCLUDE": TokenType.EXCEPT,
321            "LOGICAL": TokenType.BOOLEAN,
322            "ONLY": TokenType.ONLY,
323            "PIVOT_WIDER": TokenType.PIVOT,
324            "POSITIONAL": TokenType.POSITIONAL,
325            "SIGNED": TokenType.INT,
326            "STRING": TokenType.TEXT,
327            "SUMMARIZE": TokenType.SUMMARIZE,
328            "TIMESTAMP_S": TokenType.TIMESTAMP_S,
329            "TIMESTAMP_MS": TokenType.TIMESTAMP_MS,
330            "TIMESTAMP_NS": TokenType.TIMESTAMP_NS,
331            "TIMESTAMP_US": TokenType.TIMESTAMP,
332            "UBIGINT": TokenType.UBIGINT,
333            "UINTEGER": TokenType.UINT,
334            "USMALLINT": TokenType.USMALLINT,
335            "UTINYINT": TokenType.UTINYINT,
336            "VARCHAR": TokenType.TEXT,
337        }
338        KEYWORDS.pop("/*+")
339
340        SINGLE_TOKENS = {
341            **tokens.Tokenizer.SINGLE_TOKENS,
342            "$": TokenType.PARAMETER,
343        }
BYTE_STRINGS = [("e'", "'"), ("E'", "'")]
HEREDOC_STRINGS = ['$']
HEREDOC_TAG_IS_IDENTIFIER = True
HEREDOC_STRING_ALTERNATIVE = <TokenType.PARAMETER: 'PARAMETER'>
KEYWORDS = {'{%': <TokenType.BLOCK_START: 'BLOCK_START'>, '{%+': <TokenType.BLOCK_START: 'BLOCK_START'>, '{%-': <TokenType.BLOCK_START: 'BLOCK_START'>, '%}': <TokenType.BLOCK_END: 'BLOCK_END'>, '+%}': <TokenType.BLOCK_END: 'BLOCK_END'>, '-%}': <TokenType.BLOCK_END: 'BLOCK_END'>, '{{+': <TokenType.BLOCK_START: 'BLOCK_START'>, '{{-': <TokenType.BLOCK_START: 'BLOCK_START'>, '+}}': <TokenType.BLOCK_END: 'BLOCK_END'>, '-}}': <TokenType.BLOCK_END: 'BLOCK_END'>, '==': <TokenType.EQ: 'EQ'>, '::': <TokenType.DCOLON: 'DCOLON'>, '||': <TokenType.DPIPE: 'DPIPE'>, '>=': <TokenType.GTE: 'GTE'>, '<=': <TokenType.LTE: 'LTE'>, '<>': <TokenType.NEQ: 'NEQ'>, '!=': <TokenType.NEQ: 'NEQ'>, ':=': <TokenType.COLON_EQ: 'COLON_EQ'>, '<=>': <TokenType.NULLSAFE_EQ: 'NULLSAFE_EQ'>, '->': <TokenType.ARROW: 'ARROW'>, '->>': <TokenType.DARROW: 'DARROW'>, '=>': <TokenType.FARROW: 'FARROW'>, '#>': <TokenType.HASH_ARROW: 'HASH_ARROW'>, '#>>': <TokenType.DHASH_ARROW: 'DHASH_ARROW'>, '<->': <TokenType.LR_ARROW: 'LR_ARROW'>, '&&': <TokenType.DAMP: 'DAMP'>, '??': <TokenType.DQMARK: 'DQMARK'>, '~~~': <TokenType.GLOB: 'GLOB'>, '~~': <TokenType.LIKE: 'LIKE'>, '~~*': <TokenType.ILIKE: 'ILIKE'>, '~*': <TokenType.IRLIKE: 'IRLIKE'>, 'ALL': <TokenType.ALL: 'ALL'>, 'ALWAYS': <TokenType.ALWAYS: 'ALWAYS'>, 'AND': <TokenType.AND: 'AND'>, 'ANTI': <TokenType.ANTI: 'ANTI'>, 'ANY': <TokenType.ANY: 'ANY'>, 'ASC': <TokenType.ASC: 'ASC'>, 'AS': <TokenType.ALIAS: 'ALIAS'>, 'ASOF': <TokenType.ASOF: 'ASOF'>, 'AUTOINCREMENT': <TokenType.AUTO_INCREMENT: 'AUTO_INCREMENT'>, 'AUTO_INCREMENT': <TokenType.AUTO_INCREMENT: 'AUTO_INCREMENT'>, 'BEGIN': <TokenType.BEGIN: 'BEGIN'>, 'BETWEEN': <TokenType.BETWEEN: 'BETWEEN'>, 'CACHE': <TokenType.CACHE: 'CACHE'>, 'UNCACHE': <TokenType.UNCACHE: 'UNCACHE'>, 'CASE': <TokenType.CASE: 'CASE'>, 'CHARACTER SET': <TokenType.CHARACTER_SET: 'CHARACTER_SET'>, 'CLUSTER BY': <TokenType.CLUSTER_BY: 'CLUSTER_BY'>, 'COLLATE': <TokenType.COLLATE: 'COLLATE'>, 'COLUMN': <TokenType.COLUMN: 'COLUMN'>, 'COMMIT': <TokenType.COMMIT: 'COMMIT'>, 'CONNECT BY': <TokenType.CONNECT_BY: 'CONNECT_BY'>, 'CONSTRAINT': <TokenType.CONSTRAINT: 'CONSTRAINT'>, 'COPY': <TokenType.COPY: 'COPY'>, 'CREATE': <TokenType.CREATE: 'CREATE'>, 'CROSS': <TokenType.CROSS: 'CROSS'>, 'CUBE': <TokenType.CUBE: 'CUBE'>, 'CURRENT_DATE': <TokenType.CURRENT_DATE: 'CURRENT_DATE'>, 'CURRENT_TIME': <TokenType.CURRENT_TIME: 'CURRENT_TIME'>, 'CURRENT_TIMESTAMP': <TokenType.CURRENT_TIMESTAMP: 'CURRENT_TIMESTAMP'>, 'CURRENT_USER': <TokenType.CURRENT_USER: 'CURRENT_USER'>, 'DATABASE': <TokenType.DATABASE: 'DATABASE'>, 'DEFAULT': <TokenType.DEFAULT: 'DEFAULT'>, 'DELETE': <TokenType.DELETE: 'DELETE'>, 'DESC': <TokenType.DESC: 'DESC'>, 'DESCRIBE': <TokenType.DESCRIBE: 'DESCRIBE'>, 'DISTINCT': <TokenType.DISTINCT: 'DISTINCT'>, 'DISTRIBUTE BY': <TokenType.DISTRIBUTE_BY: 'DISTRIBUTE_BY'>, 'DIV': <TokenType.DIV: 'DIV'>, 'DROP': <TokenType.DROP: 'DROP'>, 'ELSE': <TokenType.ELSE: 'ELSE'>, 'END': <TokenType.END: 'END'>, 'ENUM': <TokenType.ENUM: 'ENUM'>, 'ESCAPE': <TokenType.ESCAPE: 'ESCAPE'>, 'EXCEPT': <TokenType.EXCEPT: 'EXCEPT'>, 'EXECUTE': <TokenType.EXECUTE: 'EXECUTE'>, 'EXISTS': <TokenType.EXISTS: 'EXISTS'>, 'FALSE': <TokenType.FALSE: 'FALSE'>, 'FETCH': <TokenType.FETCH: 'FETCH'>, 'FILTER': <TokenType.FILTER: 'FILTER'>, 'FIRST': <TokenType.FIRST: 'FIRST'>, 'FULL': <TokenType.FULL: 'FULL'>, 'FUNCTION': <TokenType.FUNCTION: 'FUNCTION'>, 'FOR': <TokenType.FOR: 'FOR'>, 'FOREIGN KEY': <TokenType.FOREIGN_KEY: 'FOREIGN_KEY'>, 'FORMAT': <TokenType.FORMAT: 'FORMAT'>, 'FROM': <TokenType.FROM: 'FROM'>, 'GEOGRAPHY': <TokenType.GEOGRAPHY: 'GEOGRAPHY'>, 'GEOMETRY': <TokenType.GEOMETRY: 'GEOMETRY'>, 'GLOB': <TokenType.GLOB: 'GLOB'>, 'GROUP BY': <TokenType.GROUP_BY: 'GROUP_BY'>, 'GROUPING SETS': <TokenType.GROUPING_SETS: 'GROUPING_SETS'>, 'HAVING': <TokenType.HAVING: 'HAVING'>, 'ILIKE': <TokenType.ILIKE: 'ILIKE'>, 'IN': <TokenType.IN: 'IN'>, 'INDEX': <TokenType.INDEX: 'INDEX'>, 'INET': <TokenType.INET: 'INET'>, 'INNER': <TokenType.INNER: 'INNER'>, 'INSERT': <TokenType.INSERT: 'INSERT'>, 'INTERVAL': <TokenType.INTERVAL: 'INTERVAL'>, 'INTERSECT': <TokenType.INTERSECT: 'INTERSECT'>, 'INTO': <TokenType.INTO: 'INTO'>, 'IS': <TokenType.IS: 'IS'>, 'ISNULL': <TokenType.ISNULL: 'ISNULL'>, 'JOIN': <TokenType.JOIN: 'JOIN'>, 'KEEP': <TokenType.KEEP: 'KEEP'>, 'KILL': <TokenType.KILL: 'KILL'>, 'LATERAL': <TokenType.LATERAL: 'LATERAL'>, 'LEFT': <TokenType.LEFT: 'LEFT'>, 'LIKE': <TokenType.LIKE: 'LIKE'>, 'LIMIT': <TokenType.LIMIT: 'LIMIT'>, 'LOAD': <TokenType.LOAD: 'LOAD'>, 'LOCK': <TokenType.LOCK: 'LOCK'>, 'MERGE': <TokenType.MERGE: 'MERGE'>, 'NAMESPACE': <TokenType.NAMESPACE: 'NAMESPACE'>, 'NATURAL': <TokenType.NATURAL: 'NATURAL'>, 'NEXT': <TokenType.NEXT: 'NEXT'>, 'NOT': <TokenType.NOT: 'NOT'>, 'NOTNULL': <TokenType.NOTNULL: 'NOTNULL'>, 'NULL': <TokenType.NULL: 'NULL'>, 'OBJECT': <TokenType.OBJECT: 'OBJECT'>, 'OFFSET': <TokenType.OFFSET: 'OFFSET'>, 'ON': <TokenType.ON: 'ON'>, 'OR': <TokenType.OR: 'OR'>, 'XOR': <TokenType.XOR: 'XOR'>, 'ORDER BY': <TokenType.ORDER_BY: 'ORDER_BY'>, 'ORDINALITY': <TokenType.ORDINALITY: 'ORDINALITY'>, 'OUTER': <TokenType.OUTER: 'OUTER'>, 'OVER': <TokenType.OVER: 'OVER'>, 'OVERLAPS': <TokenType.OVERLAPS: 'OVERLAPS'>, 'OVERWRITE': <TokenType.OVERWRITE: 'OVERWRITE'>, 'PARTITION': <TokenType.PARTITION: 'PARTITION'>, 'PARTITION BY': <TokenType.PARTITION_BY: 'PARTITION_BY'>, 'PARTITIONED BY': <TokenType.PARTITION_BY: 'PARTITION_BY'>, 'PARTITIONED_BY': <TokenType.PARTITION_BY: 'PARTITION_BY'>, 'PERCENT': <TokenType.PERCENT: 'PERCENT'>, 'PIVOT': <TokenType.PIVOT: 'PIVOT'>, 'PRAGMA': <TokenType.PRAGMA: 'PRAGMA'>, 'PRIMARY KEY': <TokenType.PRIMARY_KEY: 'PRIMARY_KEY'>, 'PROCEDURE': <TokenType.PROCEDURE: 'PROCEDURE'>, 'QUALIFY': <TokenType.QUALIFY: 'QUALIFY'>, 'RANGE': <TokenType.RANGE: 'RANGE'>, 'RECURSIVE': <TokenType.RECURSIVE: 'RECURSIVE'>, 'REGEXP': <TokenType.RLIKE: 'RLIKE'>, 'RENAME': <TokenType.RENAME: 'RENAME'>, 'REPLACE': <TokenType.REPLACE: 'REPLACE'>, 'RETURNING': <TokenType.RETURNING: 'RETURNING'>, 'REFERENCES': <TokenType.REFERENCES: 'REFERENCES'>, 'RIGHT': <TokenType.RIGHT: 'RIGHT'>, 'RLIKE': <TokenType.RLIKE: 'RLIKE'>, 'ROLLBACK': <TokenType.ROLLBACK: 'ROLLBACK'>, 'ROLLUP': <TokenType.ROLLUP: 'ROLLUP'>, 'ROW': <TokenType.ROW: 'ROW'>, 'ROWS': <TokenType.ROWS: 'ROWS'>, 'SCHEMA': <TokenType.SCHEMA: 'SCHEMA'>, 'SELECT': <TokenType.SELECT: 'SELECT'>, 'SEMI': <TokenType.SEMI: 'SEMI'>, 'SET': <TokenType.SET: 'SET'>, 'SETTINGS': <TokenType.SETTINGS: 'SETTINGS'>, 'SHOW': <TokenType.SHOW: 'SHOW'>, 'SIMILAR TO': <TokenType.SIMILAR_TO: 'SIMILAR_TO'>, 'SOME': <TokenType.SOME: 'SOME'>, 'SORT BY': <TokenType.SORT_BY: 'SORT_BY'>, 'START WITH': <TokenType.START_WITH: 'START_WITH'>, 'STRAIGHT_JOIN': <TokenType.STRAIGHT_JOIN: 'STRAIGHT_JOIN'>, 'TABLE': <TokenType.TABLE: 'TABLE'>, 'TABLESAMPLE': <TokenType.TABLE_SAMPLE: 'TABLE_SAMPLE'>, 'TEMP': <TokenType.TEMPORARY: 'TEMPORARY'>, 'TEMPORARY': <TokenType.TEMPORARY: 'TEMPORARY'>, 'THEN': <TokenType.THEN: 'THEN'>, 'TRUE': <TokenType.TRUE: 'TRUE'>, 'TRUNCATE': <TokenType.TRUNCATE: 'TRUNCATE'>, 'UNION': <TokenType.UNION: 'UNION'>, 'UNKNOWN': <TokenType.UNKNOWN: 'UNKNOWN'>, 'UNNEST': <TokenType.UNNEST: 'UNNEST'>, 'UNPIVOT': <TokenType.UNPIVOT: 'UNPIVOT'>, 'UPDATE': <TokenType.UPDATE: 'UPDATE'>, 'USE': <TokenType.USE: 'USE'>, 'USING': <TokenType.USING: 'USING'>, 'UUID': <TokenType.UUID: 'UUID'>, 'VALUES': <TokenType.VALUES: 'VALUES'>, 'VIEW': <TokenType.VIEW: 'VIEW'>, 'VOLATILE': <TokenType.VOLATILE: 'VOLATILE'>, 'WHEN': <TokenType.WHEN: 'WHEN'>, 'WHERE': <TokenType.WHERE: 'WHERE'>, 'WINDOW': <TokenType.WINDOW: 'WINDOW'>, 'WITH': <TokenType.WITH: 'WITH'>, 'APPLY': <TokenType.APPLY: 'APPLY'>, 'ARRAY': <TokenType.ARRAY: 'ARRAY'>, 'BIT': <TokenType.BIT: 'BIT'>, 'BOOL': <TokenType.BOOLEAN: 'BOOLEAN'>, 'BOOLEAN': <TokenType.BOOLEAN: 'BOOLEAN'>, 'BYTE': <TokenType.TINYINT: 'TINYINT'>, 'MEDIUMINT': <TokenType.MEDIUMINT: 'MEDIUMINT'>, 'INT1': <TokenType.TINYINT: 'TINYINT'>, 'TINYINT': <TokenType.TINYINT: 'TINYINT'>, 'INT16': <TokenType.SMALLINT: 'SMALLINT'>, 'SHORT': <TokenType.SMALLINT: 'SMALLINT'>, 'SMALLINT': <TokenType.SMALLINT: 'SMALLINT'>, 'INT128': <TokenType.INT128: 'INT128'>, 'HUGEINT': <TokenType.INT128: 'INT128'>, 'UHUGEINT': <TokenType.UINT128: 'UINT128'>, 'INT2': <TokenType.SMALLINT: 'SMALLINT'>, 'INTEGER': <TokenType.INT: 'INT'>, 'INT': <TokenType.INT: 'INT'>, 'INT4': <TokenType.INT: 'INT'>, 'INT32': <TokenType.INT: 'INT'>, 'INT64': <TokenType.BIGINT: 'BIGINT'>, 'LONG': <TokenType.BIGINT: 'BIGINT'>, 'BIGINT': <TokenType.BIGINT: 'BIGINT'>, 'INT8': <TokenType.TINYINT: 'TINYINT'>, 'UINT': <TokenType.UINT: 'UINT'>, 'DEC': <TokenType.DECIMAL: 'DECIMAL'>, 'DECIMAL': <TokenType.DECIMAL: 'DECIMAL'>, 'DECIMAL32': <TokenType.DECIMAL32: 'DECIMAL32'>, 'DECIMAL64': <TokenType.DECIMAL64: 'DECIMAL64'>, 'DECIMAL128': <TokenType.DECIMAL128: 'DECIMAL128'>, 'DECIMAL256': <TokenType.DECIMAL256: 'DECIMAL256'>, 'BIGDECIMAL': <TokenType.BIGDECIMAL: 'BIGDECIMAL'>, 'BIGNUMERIC': <TokenType.BIGDECIMAL: 'BIGDECIMAL'>, 'LIST': <TokenType.LIST: 'LIST'>, 'MAP': <TokenType.MAP: 'MAP'>, 'NULLABLE': <TokenType.NULLABLE: 'NULLABLE'>, 'NUMBER': <TokenType.DECIMAL: 'DECIMAL'>, 'NUMERIC': <TokenType.DECIMAL: 'DECIMAL'>, 'FIXED': <TokenType.DECIMAL: 'DECIMAL'>, 'REAL': <TokenType.FLOAT: 'FLOAT'>, 'FLOAT': <TokenType.FLOAT: 'FLOAT'>, 'FLOAT4': <TokenType.FLOAT: 'FLOAT'>, 'FLOAT8': <TokenType.DOUBLE: 'DOUBLE'>, 'DOUBLE': <TokenType.DOUBLE: 'DOUBLE'>, 'DOUBLE PRECISION': <TokenType.DOUBLE: 'DOUBLE'>, 'JSON': <TokenType.JSON: 'JSON'>, 'JSONB': <TokenType.JSONB: 'JSONB'>, 'CHAR': <TokenType.TEXT: 'TEXT'>, 'CHARACTER': <TokenType.CHAR: 'CHAR'>, 'NCHAR': <TokenType.NCHAR: 'NCHAR'>, 'VARCHAR': <TokenType.TEXT: 'TEXT'>, 'VARCHAR2': <TokenType.VARCHAR: 'VARCHAR'>, 'NVARCHAR': <TokenType.NVARCHAR: 'NVARCHAR'>, 'NVARCHAR2': <TokenType.NVARCHAR: 'NVARCHAR'>, 'BPCHAR': <TokenType.TEXT: 'TEXT'>, 'STR': <TokenType.TEXT: 'TEXT'>, 'STRING': <TokenType.TEXT: 'TEXT'>, 'TEXT': <TokenType.TEXT: 'TEXT'>, 'LONGTEXT': <TokenType.LONGTEXT: 'LONGTEXT'>, 'MEDIUMTEXT': <TokenType.MEDIUMTEXT: 'MEDIUMTEXT'>, 'TINYTEXT': <TokenType.TINYTEXT: 'TINYTEXT'>, 'CLOB': <TokenType.TEXT: 'TEXT'>, 'LONGVARCHAR': <TokenType.TEXT: 'TEXT'>, 'BINARY': <TokenType.VARBINARY: 'VARBINARY'>, 'BLOB': <TokenType.VARBINARY: 'VARBINARY'>, 'LONGBLOB': <TokenType.LONGBLOB: 'LONGBLOB'>, 'MEDIUMBLOB': <TokenType.MEDIUMBLOB: 'MEDIUMBLOB'>, 'TINYBLOB': <TokenType.TINYBLOB: 'TINYBLOB'>, 'BYTEA': <TokenType.VARBINARY: 'VARBINARY'>, 'VARBINARY': <TokenType.VARBINARY: 'VARBINARY'>, 'TIME': <TokenType.TIME: 'TIME'>, 'TIMETZ': <TokenType.TIMETZ: 'TIMETZ'>, 'TIMESTAMP': <TokenType.TIMESTAMP: 'TIMESTAMP'>, 'TIMESTAMPTZ': <TokenType.TIMESTAMPTZ: 'TIMESTAMPTZ'>, 'TIMESTAMPLTZ': <TokenType.TIMESTAMPLTZ: 'TIMESTAMPLTZ'>, 'TIMESTAMP_LTZ': <TokenType.TIMESTAMPLTZ: 'TIMESTAMPLTZ'>, 'TIMESTAMPNTZ': <TokenType.TIMESTAMPNTZ: 'TIMESTAMPNTZ'>, 'TIMESTAMP_NTZ': <TokenType.TIMESTAMPNTZ: 'TIMESTAMPNTZ'>, 'DATE': <TokenType.DATE: 'DATE'>, 'DATETIME': <TokenType.DATETIME: 'DATETIME'>, 'INT4RANGE': <TokenType.INT4RANGE: 'INT4RANGE'>, 'INT4MULTIRANGE': <TokenType.INT4MULTIRANGE: 'INT4MULTIRANGE'>, 'INT8RANGE': <TokenType.INT8RANGE: 'INT8RANGE'>, 'INT8MULTIRANGE': <TokenType.INT8MULTIRANGE: 'INT8MULTIRANGE'>, 'NUMRANGE': <TokenType.NUMRANGE: 'NUMRANGE'>, 'NUMMULTIRANGE': <TokenType.NUMMULTIRANGE: 'NUMMULTIRANGE'>, 'TSRANGE': <TokenType.TSRANGE: 'TSRANGE'>, 'TSMULTIRANGE': <TokenType.TSMULTIRANGE: 'TSMULTIRANGE'>, 'TSTZRANGE': <TokenType.TSTZRANGE: 'TSTZRANGE'>, 'TSTZMULTIRANGE': <TokenType.TSTZMULTIRANGE: 'TSTZMULTIRANGE'>, 'DATERANGE': <TokenType.DATERANGE: 'DATERANGE'>, 'DATEMULTIRANGE': <TokenType.DATEMULTIRANGE: 'DATEMULTIRANGE'>, 'UNIQUE': <TokenType.UNIQUE: 'UNIQUE'>, 'VECTOR': <TokenType.VECTOR: 'VECTOR'>, 'STRUCT': <TokenType.STRUCT: 'STRUCT'>, 'SEQUENCE': <TokenType.SEQUENCE: 'SEQUENCE'>, 'VARIANT': <TokenType.VARIANT: 'VARIANT'>, 'ALTER': <TokenType.ALTER: 'ALTER'>, 'ANALYZE': <TokenType.ANALYZE: 'ANALYZE'>, 'CALL': <TokenType.COMMAND: 'COMMAND'>, 'COMMENT': <TokenType.COMMENT: 'COMMENT'>, 'EXPLAIN': <TokenType.COMMAND: 'COMMAND'>, 'GRANT': <TokenType.GRANT: 'GRANT'>, 'OPTIMIZE': <TokenType.COMMAND: 'COMMAND'>, 'PREPARE': <TokenType.COMMAND: 'COMMAND'>, 'VACUUM': <TokenType.COMMAND: 'COMMAND'>, 'USER-DEFINED': <TokenType.USERDEFINED: 'USERDEFINED'>, 'FOR VERSION': <TokenType.VERSION_SNAPSHOT: 'VERSION_SNAPSHOT'>, 'FOR TIMESTAMP': <TokenType.TIMESTAMP_SNAPSHOT: 'TIMESTAMP_SNAPSHOT'>, '//': <TokenType.DIV: 'DIV'>, '**': <TokenType.DSTAR: 'DSTAR'>, '^@': <TokenType.CARET_AT: 'CARET_AT'>, '@>': <TokenType.AT_GT: 'AT_GT'>, '<@': <TokenType.LT_AT: 'LT_AT'>, 'ATTACH': <TokenType.ATTACH: 'ATTACH'>, 'BITSTRING': <TokenType.BIT: 'BIT'>, 'CHARACTER VARYING': <TokenType.TEXT: 'TEXT'>, 'DETACH': <TokenType.DETACH: 'DETACH'>, 'EXCLUDE': <TokenType.EXCEPT: 'EXCEPT'>, 'LOGICAL': <TokenType.BOOLEAN: 'BOOLEAN'>, 'ONLY': <TokenType.ONLY: 'ONLY'>, 'PIVOT_WIDER': <TokenType.PIVOT: 'PIVOT'>, 'POSITIONAL': <TokenType.POSITIONAL: 'POSITIONAL'>, 'SIGNED': <TokenType.INT: 'INT'>, 'SUMMARIZE': <TokenType.SUMMARIZE: 'SUMMARIZE'>, 'TIMESTAMP_S': <TokenType.TIMESTAMP_S: 'TIMESTAMP_S'>, 'TIMESTAMP_MS': <TokenType.TIMESTAMP_MS: 'TIMESTAMP_MS'>, 'TIMESTAMP_NS': <TokenType.TIMESTAMP_NS: 'TIMESTAMP_NS'>, 'TIMESTAMP_US': <TokenType.TIMESTAMP: 'TIMESTAMP'>, 'UBIGINT': <TokenType.UBIGINT: 'UBIGINT'>, 'UINTEGER': <TokenType.UINT: 'UINT'>, 'USMALLINT': <TokenType.USMALLINT: 'USMALLINT'>, 'UTINYINT': <TokenType.UTINYINT: 'UTINYINT'>}
SINGLE_TOKENS = {'(': <TokenType.L_PAREN: 'L_PAREN'>, ')': <TokenType.R_PAREN: 'R_PAREN'>, '[': <TokenType.L_BRACKET: 'L_BRACKET'>, ']': <TokenType.R_BRACKET: 'R_BRACKET'>, '{': <TokenType.L_BRACE: 'L_BRACE'>, '}': <TokenType.R_BRACE: 'R_BRACE'>, '&': <TokenType.AMP: 'AMP'>, '^': <TokenType.CARET: 'CARET'>, ':': <TokenType.COLON: 'COLON'>, ',': <TokenType.COMMA: 'COMMA'>, '.': <TokenType.DOT: 'DOT'>, '-': <TokenType.DASH: 'DASH'>, '=': <TokenType.EQ: 'EQ'>, '>': <TokenType.GT: 'GT'>, '<': <TokenType.LT: 'LT'>, '%': <TokenType.MOD: 'MOD'>, '!': <TokenType.NOT: 'NOT'>, '|': <TokenType.PIPE: 'PIPE'>, '+': <TokenType.PLUS: 'PLUS'>, ';': <TokenType.SEMICOLON: 'SEMICOLON'>, '/': <TokenType.SLASH: 'SLASH'>, '\\': <TokenType.BACKSLASH: 'BACKSLASH'>, '*': <TokenType.STAR: 'STAR'>, '~': <TokenType.TILDA: 'TILDA'>, '?': <TokenType.PLACEHOLDER: 'PLACEHOLDER'>, '@': <TokenType.PARAMETER: 'PARAMETER'>, '#': <TokenType.HASH: 'HASH'>, "'": <TokenType.UNKNOWN: 'UNKNOWN'>, '`': <TokenType.UNKNOWN: 'UNKNOWN'>, '"': <TokenType.UNKNOWN: 'UNKNOWN'>, '$': <TokenType.PARAMETER: 'PARAMETER'>}
class DuckDB.Parser(sqlglot.parser.Parser):
345    class Parser(parser.Parser):
346        BITWISE = {
347            **parser.Parser.BITWISE,
348            TokenType.TILDA: exp.RegexpLike,
349        }
350        BITWISE.pop(TokenType.CARET)
351
352        RANGE_PARSERS = {
353            **parser.Parser.RANGE_PARSERS,
354            TokenType.DAMP: binary_range_parser(exp.ArrayOverlaps),
355            TokenType.CARET_AT: binary_range_parser(exp.StartsWith),
356        }
357
358        EXPONENT = {
359            **parser.Parser.EXPONENT,
360            TokenType.CARET: exp.Pow,
361            TokenType.DSTAR: exp.Pow,
362        }
363
364        FUNCTIONS_WITH_ALIASED_ARGS = {*parser.Parser.FUNCTIONS_WITH_ALIASED_ARGS, "STRUCT_PACK"}
365
366        FUNCTIONS = {
367            **parser.Parser.FUNCTIONS,
368            "ARRAY_REVERSE_SORT": _build_sort_array_desc,
369            "ARRAY_SORT": exp.SortArray.from_arg_list,
370            "DATEDIFF": _build_date_diff,
371            "DATE_DIFF": _build_date_diff,
372            "DATE_TRUNC": date_trunc_to_time,
373            "DATETRUNC": date_trunc_to_time,
374            "DECODE": lambda args: exp.Decode(
375                this=seq_get(args, 0), charset=exp.Literal.string("utf-8")
376            ),
377            "ENCODE": lambda args: exp.Encode(
378                this=seq_get(args, 0), charset=exp.Literal.string("utf-8")
379            ),
380            "EPOCH": exp.TimeToUnix.from_arg_list,
381            "EPOCH_MS": lambda args: exp.UnixToTime(
382                this=seq_get(args, 0), scale=exp.UnixToTime.MILLIS
383            ),
384            "JSON": exp.ParseJSON.from_arg_list,
385            "JSON_EXTRACT_PATH": parser.build_extract_json_with_path(exp.JSONExtract),
386            "JSON_EXTRACT_STRING": parser.build_extract_json_with_path(exp.JSONExtractScalar),
387            "LIST_HAS": exp.ArrayContains.from_arg_list,
388            "LIST_REVERSE_SORT": _build_sort_array_desc,
389            "LIST_SORT": exp.SortArray.from_arg_list,
390            "LIST_VALUE": lambda args: exp.Array(expressions=args),
391            "MAKE_TIME": exp.TimeFromParts.from_arg_list,
392            "MAKE_TIMESTAMP": _build_make_timestamp,
393            "QUANTILE_CONT": exp.PercentileCont.from_arg_list,
394            "QUANTILE_DISC": exp.PercentileDisc.from_arg_list,
395            "REGEXP_EXTRACT": build_regexp_extract(exp.RegexpExtract),
396            "REGEXP_EXTRACT_ALL": build_regexp_extract(exp.RegexpExtractAll),
397            "REGEXP_MATCHES": exp.RegexpLike.from_arg_list,
398            "REGEXP_REPLACE": lambda args: exp.RegexpReplace(
399                this=seq_get(args, 0),
400                expression=seq_get(args, 1),
401                replacement=seq_get(args, 2),
402                modifiers=seq_get(args, 3),
403            ),
404            "STRFTIME": build_formatted_time(exp.TimeToStr, "duckdb"),
405            "STRING_SPLIT": exp.Split.from_arg_list,
406            "STRING_SPLIT_REGEX": exp.RegexpSplit.from_arg_list,
407            "STRING_TO_ARRAY": exp.Split.from_arg_list,
408            "STRPTIME": build_formatted_time(exp.StrToTime, "duckdb"),
409            "STRUCT_PACK": exp.Struct.from_arg_list,
410            "STR_SPLIT": exp.Split.from_arg_list,
411            "STR_SPLIT_REGEX": exp.RegexpSplit.from_arg_list,
412            "TO_TIMESTAMP": exp.UnixToTime.from_arg_list,
413            "UNNEST": exp.Explode.from_arg_list,
414            "XOR": binary_from_function(exp.BitwiseXor),
415            "GENERATE_SERIES": _build_generate_series(),
416            "RANGE": _build_generate_series(end_exclusive=True),
417            "EDITDIST3": exp.Levenshtein.from_arg_list,
418        }
419
420        FUNCTIONS.pop("DATE_SUB")
421        FUNCTIONS.pop("GLOB")
422
423        FUNCTION_PARSERS = parser.Parser.FUNCTION_PARSERS.copy()
424        FUNCTION_PARSERS.pop("DECODE")
425
426        NO_PAREN_FUNCTION_PARSERS = {
427            **parser.Parser.NO_PAREN_FUNCTION_PARSERS,
428            "MAP": lambda self: self._parse_map(),
429        }
430
431        TABLE_ALIAS_TOKENS = parser.Parser.TABLE_ALIAS_TOKENS - {
432            TokenType.SEMI,
433            TokenType.ANTI,
434        }
435
436        PLACEHOLDER_PARSERS = {
437            **parser.Parser.PLACEHOLDER_PARSERS,
438            TokenType.PARAMETER: lambda self: (
439                self.expression(exp.Placeholder, this=self._prev.text)
440                if self._match(TokenType.NUMBER) or self._match_set(self.ID_VAR_TOKENS)
441                else None
442            ),
443        }
444
445        TYPE_CONVERTERS = {
446            # https://duckdb.org/docs/sql/data_types/numeric
447            exp.DataType.Type.DECIMAL: build_default_decimal_type(precision=18, scale=3),
448            # https://duckdb.org/docs/sql/data_types/text
449            exp.DataType.Type.TEXT: lambda dtype: exp.DataType.build("TEXT"),
450        }
451
452        STATEMENT_PARSERS = {
453            **parser.Parser.STATEMENT_PARSERS,
454            TokenType.ATTACH: lambda self: self._parse_attach_detach(),
455            TokenType.DETACH: lambda self: self._parse_attach_detach(is_attach=False),
456        }
457
458        def _parse_table_sample(self, as_modifier: bool = False) -> t.Optional[exp.TableSample]:
459            # https://duckdb.org/docs/sql/samples.html
460            sample = super()._parse_table_sample(as_modifier=as_modifier)
461            if sample and not sample.args.get("method"):
462                if sample.args.get("size"):
463                    sample.set("method", exp.var("RESERVOIR"))
464                else:
465                    sample.set("method", exp.var("SYSTEM"))
466
467            return sample
468
469        def _parse_bracket(
470            self, this: t.Optional[exp.Expression] = None
471        ) -> t.Optional[exp.Expression]:
472            bracket = super()._parse_bracket(this)
473            if isinstance(bracket, exp.Bracket):
474                bracket.set("returns_list_for_maps", True)
475
476            return bracket
477
478        def _parse_map(self) -> exp.ToMap | exp.Map:
479            if self._match(TokenType.L_BRACE, advance=False):
480                return self.expression(exp.ToMap, this=self._parse_bracket())
481
482            args = self._parse_wrapped_csv(self._parse_assignment)
483            return self.expression(exp.Map, keys=seq_get(args, 0), values=seq_get(args, 1))
484
485        def _parse_struct_types(self, type_required: bool = False) -> t.Optional[exp.Expression]:
486            return self._parse_field_def()
487
488        def _pivot_column_names(self, aggregations: t.List[exp.Expression]) -> t.List[str]:
489            if len(aggregations) == 1:
490                return super()._pivot_column_names(aggregations)
491            return pivot_column_names(aggregations, dialect="duckdb")
492
493        def _parse_attach_detach(self, is_attach=True) -> exp.Attach | exp.Detach:
494            def _parse_attach_option() -> exp.AttachOption:
495                return self.expression(
496                    exp.AttachOption,
497                    this=self._parse_var(any_token=True),
498                    expression=self._parse_field(any_token=True),
499                )
500
501            self._match(TokenType.DATABASE)
502            exists = self._parse_exists(not_=is_attach)
503            this = self._parse_alias(self._parse_primary_or_var(), explicit=True)
504
505            if self._match(TokenType.L_PAREN, advance=False):
506                expressions = self._parse_wrapped_csv(_parse_attach_option)
507            else:
508                expressions = None
509
510            return (
511                self.expression(exp.Attach, this=this, exists=exists, expressions=expressions)
512                if is_attach
513                else self.expression(exp.Detach, this=this, exists=exists)
514            )

Parser consumes a list of tokens produced by the Tokenizer and produces a parsed syntax tree.

Arguments:
  • error_level: The desired error level. Default: ErrorLevel.IMMEDIATE
  • error_message_context: The amount of context to capture from a query string when displaying the error message (in number of characters). Default: 100
  • max_errors: Maximum number of error messages to include in a raised ParseError. This is only relevant if error_level is ErrorLevel.RAISE. Default: 3
BITWISE = {<TokenType.AMP: 'AMP'>: <class 'sqlglot.expressions.BitwiseAnd'>, <TokenType.PIPE: 'PIPE'>: <class 'sqlglot.expressions.BitwiseOr'>, <TokenType.TILDA: 'TILDA'>: <class 'sqlglot.expressions.RegexpLike'>}
RANGE_PARSERS = {<TokenType.AT_GT: 'AT_GT'>: <function binary_range_parser.<locals>._parse_binary_range>, <TokenType.BETWEEN: 'BETWEEN'>: <function Parser.<lambda>>, <TokenType.GLOB: 'GLOB'>: <function binary_range_parser.<locals>._parse_binary_range>, <TokenType.ILIKE: 'ILIKE'>: <function binary_range_parser.<locals>._parse_binary_range>, <TokenType.IN: 'IN'>: <function Parser.<lambda>>, <TokenType.IRLIKE: 'IRLIKE'>: <function binary_range_parser.<locals>._parse_binary_range>, <TokenType.IS: 'IS'>: <function Parser.<lambda>>, <TokenType.LIKE: 'LIKE'>: <function binary_range_parser.<locals>._parse_binary_range>, <TokenType.LT_AT: 'LT_AT'>: <function binary_range_parser.<locals>._parse_binary_range>, <TokenType.OVERLAPS: 'OVERLAPS'>: <function binary_range_parser.<locals>._parse_binary_range>, <TokenType.RLIKE: 'RLIKE'>: <function binary_range_parser.<locals>._parse_binary_range>, <TokenType.SIMILAR_TO: 'SIMILAR_TO'>: <function binary_range_parser.<locals>._parse_binary_range>, <TokenType.FOR: 'FOR'>: <function Parser.<lambda>>, <TokenType.DAMP: 'DAMP'>: <function binary_range_parser.<locals>._parse_binary_range>, <TokenType.CARET_AT: 'CARET_AT'>: <function binary_range_parser.<locals>._parse_binary_range>}
EXPONENT = {<TokenType.CARET: 'CARET'>: <class 'sqlglot.expressions.Pow'>, <TokenType.DSTAR: 'DSTAR'>: <class 'sqlglot.expressions.Pow'>}
FUNCTIONS_WITH_ALIASED_ARGS = {'STRUCT_PACK', 'STRUCT'}
FUNCTIONS = {'ABS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Abs'>>, 'ADD_MONTHS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.AddMonths'>>, 'ANONYMOUS_AGG_FUNC': <bound method Func.from_arg_list of <class 'sqlglot.expressions.AnonymousAggFunc'>>, 'ANY_VALUE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.AnyValue'>>, 'APPLY': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Apply'>>, 'APPROX_DISTINCT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ApproxDistinct'>>, 'APPROX_COUNT_DISTINCT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ApproxDistinct'>>, 'APPROX_QUANTILE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ApproxQuantile'>>, 'APPROX_TOP_K': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ApproxTopK'>>, 'ARG_MAX': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArgMax'>>, 'ARGMAX': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArgMax'>>, 'MAX_BY': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArgMax'>>, 'ARG_MIN': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArgMin'>>, 'ARGMIN': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArgMin'>>, 'MIN_BY': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArgMin'>>, 'ARRAY': <function Parser.<lambda>>, 'ARRAY_AGG': <function Parser.<lambda>>, 'ARRAY_ALL': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArrayAll'>>, 'ARRAY_ANY': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArrayAny'>>, 'ARRAY_CONCAT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArrayConcat'>>, 'ARRAY_CAT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArrayConcat'>>, 'ARRAY_CONSTRUCT_COMPACT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArrayConstructCompact'>>, 'ARRAY_CONTAINS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArrayContains'>>, 'ARRAY_HAS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArrayContains'>>, 'ARRAY_CONTAINS_ALL': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArrayContainsAll'>>, 'ARRAY_HAS_ALL': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArrayContainsAll'>>, 'FILTER': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArrayFilter'>>, 'ARRAY_FILTER': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArrayFilter'>>, 'ARRAY_OVERLAPS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArrayOverlaps'>>, 'ARRAY_SIZE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArraySize'>>, 'ARRAY_LENGTH': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArraySize'>>, 'ARRAY_SORT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.SortArray'>>, 'ARRAY_SUM': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArraySum'>>, 'ARRAY_TO_STRING': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArrayToString'>>, 'ARRAY_JOIN': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArrayToString'>>, 'ARRAY_UNION_AGG': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArrayUnionAgg'>>, 'ARRAY_UNIQUE_AGG': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArrayUniqueAgg'>>, 'AVG': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Avg'>>, 'CASE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Case'>>, 'CAST': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Cast'>>, 'CAST_TO_STR_TYPE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.CastToStrType'>>, 'CBRT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Cbrt'>>, 'CEIL': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Ceil'>>, 'CEILING': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Ceil'>>, 'CHR': <function Parser.<lambda>>, 'CHAR': <function Parser.<lambda>>, 'COALESCE': <function build_coalesce>, 'IFNULL': <function build_coalesce>, 'NVL': <function build_coalesce>, 'COLLATE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Collate'>>, 'COLUMNS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Columns'>>, 'COMBINED_AGG_FUNC': <bound method Func.from_arg_list of <class 'sqlglot.expressions.CombinedAggFunc'>>, 'COMBINED_PARAMETERIZED_AGG': <bound method Func.from_arg_list of <class 'sqlglot.expressions.CombinedParameterizedAgg'>>, 'CONCAT': <function Parser.<lambda>>, 'CONCAT_WS': <function Parser.<lambda>>, 'CONNECT_BY_ROOT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ConnectByRoot'>>, 'CONTAINS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Contains'>>, 'CONVERT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Convert'>>, 'CONVERT_TIMEZONE': <function build_convert_timezone>, 'CORR': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Corr'>>, 'COUNT': <function Parser.<lambda>>, 'COUNT_IF': <bound method Func.from_arg_list of <class 'sqlglot.expressions.CountIf'>>, 'COUNTIF': <bound method Func.from_arg_list of <class 'sqlglot.expressions.CountIf'>>, 'COVAR_POP': <bound method Func.from_arg_list of <class 'sqlglot.expressions.CovarPop'>>, 'COVAR_SAMP': <bound method Func.from_arg_list of <class 'sqlglot.expressions.CovarSamp'>>, 'CURRENT_DATE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.CurrentDate'>>, 'CURRENT_DATETIME': <bound method Func.from_arg_list of <class 'sqlglot.expressions.CurrentDatetime'>>, 'CURRENT_TIME': <bound method Func.from_arg_list of <class 'sqlglot.expressions.CurrentTime'>>, 'CURRENT_TIMESTAMP': <bound method Func.from_arg_list of <class 'sqlglot.expressions.CurrentTimestamp'>>, 'CURRENT_USER': <bound method Func.from_arg_list of <class 'sqlglot.expressions.CurrentUser'>>, 'DATE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Date'>>, 'DATE_ADD': <bound method Func.from_arg_list of <class 'sqlglot.expressions.DateAdd'>>, 'DATEDIFF': <function _build_date_diff>, 'DATE_DIFF': <function _build_date_diff>, 'DATE_FROM_PARTS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.DateFromParts'>>, 'DATEFROMPARTS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.DateFromParts'>>, 'DATE_STR_TO_DATE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.DateStrToDate'>>, 'DATE_TO_DATE_STR': <function Parser.<lambda>>, 'DATE_TO_DI': <bound method Func.from_arg_list of <class 'sqlglot.expressions.DateToDi'>>, 'DATE_TRUNC': <function date_trunc_to_time>, 'DATETIME': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Datetime'>>, 'DATETIME_ADD': <bound method Func.from_arg_list of <class 'sqlglot.expressions.DatetimeAdd'>>, 'DATETIME_DIFF': <bound method Func.from_arg_list of <class 'sqlglot.expressions.DatetimeDiff'>>, 'DATETIME_SUB': <bound method Func.from_arg_list of <class 'sqlglot.expressions.DatetimeSub'>>, 'DATETIME_TRUNC': <bound method Func.from_arg_list of <class 'sqlglot.expressions.DatetimeTrunc'>>, 'DAY': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Day'>>, 'DAY_OF_MONTH': <bound method Func.from_arg_list of <class 'sqlglot.expressions.DayOfMonth'>>, 'DAYOFMONTH': <bound method Func.from_arg_list of <class 'sqlglot.expressions.DayOfMonth'>>, 'DAY_OF_WEEK': <bound method Func.from_arg_list of <class 'sqlglot.expressions.DayOfWeek'>>, 'DAYOFWEEK': <bound method Func.from_arg_list of <class 'sqlglot.expressions.DayOfWeek'>>, 'DAYOFWEEK_ISO': <bound method Func.from_arg_list of <class 'sqlglot.expressions.DayOfWeekIso'>>, 'ISODOW': <bound method Func.from_arg_list of <class 'sqlglot.expressions.DayOfWeekIso'>>, 'DAY_OF_YEAR': <bound method Func.from_arg_list of <class 'sqlglot.expressions.DayOfYear'>>, 'DAYOFYEAR': <bound method Func.from_arg_list of <class 'sqlglot.expressions.DayOfYear'>>, 'DECODE': <function DuckDB.Parser.<lambda>>, 'DI_TO_DATE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.DiToDate'>>, 'ENCODE': <function DuckDB.Parser.<lambda>>, 'EXISTS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Exists'>>, 'EXP': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Exp'>>, 'EXPLODE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Explode'>>, 'EXPLODE_OUTER': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ExplodeOuter'>>, 'EXPLODING_GENERATE_SERIES': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ExplodingGenerateSeries'>>, 'EXTRACT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Extract'>>, 'FEATURES_AT_TIME': <bound method Func.from_arg_list of <class 'sqlglot.expressions.FeaturesAtTime'>>, 'FIRST': <bound method Func.from_arg_list of <class 'sqlglot.expressions.First'>>, 'FIRST_VALUE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.FirstValue'>>, 'FLATTEN': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Flatten'>>, 'FLOOR': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Floor'>>, 'FROM_BASE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.FromBase'>>, 'FROM_BASE64': <bound method Func.from_arg_list of <class 'sqlglot.expressions.FromBase64'>>, 'FROM_ISO8601_TIMESTAMP': <bound method Func.from_arg_list of <class 'sqlglot.expressions.FromISO8601Timestamp'>>, 'GAP_FILL': <bound method Func.from_arg_list of <class 'sqlglot.expressions.GapFill'>>, 'GENERATE_DATE_ARRAY': <function Parser.<lambda>>, 'GENERATE_SERIES': <function _build_generate_series.<locals>._builder>, 'GENERATE_TIMESTAMP_ARRAY': <bound method Func.from_arg_list of <class 'sqlglot.expressions.GenerateTimestampArray'>>, 'GREATEST': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Greatest'>>, 'GROUP_CONCAT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.GroupConcat'>>, 'HEX': <function build_hex>, 'HLL': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Hll'>>, 'IF': <bound method Func.from_arg_list of <class 'sqlglot.expressions.If'>>, 'IIF': <bound method Func.from_arg_list of <class 'sqlglot.expressions.If'>>, 'INITCAP': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Initcap'>>, 'INLINE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Inline'>>, 'INT64': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Int64'>>, 'IS_ASCII': <bound method Func.from_arg_list of <class 'sqlglot.expressions.IsAscii'>>, 'IS_INF': <bound method Func.from_arg_list of <class 'sqlglot.expressions.IsInf'>>, 'ISINF': <bound method Func.from_arg_list of <class 'sqlglot.expressions.IsInf'>>, 'IS_NAN': <bound method Func.from_arg_list of <class 'sqlglot.expressions.IsNan'>>, 'ISNAN': <bound method Func.from_arg_list of <class 'sqlglot.expressions.IsNan'>>, 'J_S_O_N_ARRAY': <bound method Func.from_arg_list of <class 'sqlglot.expressions.JSONArray'>>, 'J_S_O_N_ARRAY_AGG': <bound method Func.from_arg_list of <class 'sqlglot.expressions.JSONArrayAgg'>>, 'JSON_ARRAY_CONTAINS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.JSONArrayContains'>>, 'JSONB_CONTAINS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.JSONBContains'>>, 'JSONB_EXISTS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.JSONBExists'>>, 'JSONB_EXTRACT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.JSONBExtract'>>, 'JSONB_EXTRACT_SCALAR': <bound method Func.from_arg_list of <class 'sqlglot.expressions.JSONBExtractScalar'>>, 'J_S_O_N_EXISTS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.JSONExists'>>, 'JSON_EXTRACT': <function build_extract_json_with_path.<locals>._builder>, 'JSON_EXTRACT_ARRAY': <bound method Func.from_arg_list of <class 'sqlglot.expressions.JSONExtractArray'>>, 'JSON_EXTRACT_SCALAR': <function build_extract_json_with_path.<locals>._builder>, 'JSON_FORMAT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.JSONFormat'>>, 'J_S_O_N_OBJECT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.JSONObject'>>, 'J_S_O_N_OBJECT_AGG': <bound method Func.from_arg_list of <class 'sqlglot.expressions.JSONObjectAgg'>>, 'J_S_O_N_TABLE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.JSONTable'>>, 'J_S_O_N_VALUE_ARRAY': <bound method Func.from_arg_list of <class 'sqlglot.expressions.JSONValueArray'>>, 'LAG': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Lag'>>, 'LAST': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Last'>>, 'LAST_DAY': <bound method Func.from_arg_list of <class 'sqlglot.expressions.LastDay'>>, 'LAST_DAY_OF_MONTH': <bound method Func.from_arg_list of <class 'sqlglot.expressions.LastDay'>>, 'LAST_VALUE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.LastValue'>>, 'LEAD': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Lead'>>, 'LEAST': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Least'>>, 'LEFT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Left'>>, 'LENGTH': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Length'>>, 'LEN': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Length'>>, 'CHAR_LENGTH': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Length'>>, 'CHARACTER_LENGTH': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Length'>>, 'LEVENSHTEIN': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Levenshtein'>>, 'LIST': <bound method Func.from_arg_list of <class 'sqlglot.expressions.List'>>, 'LN': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Ln'>>, 'LOG': <function build_logarithm>, 'LOGICAL_AND': <bound method Func.from_arg_list of <class 'sqlglot.expressions.LogicalAnd'>>, 'BOOL_AND': <bound method Func.from_arg_list of <class 'sqlglot.expressions.LogicalAnd'>>, 'BOOLAND_AGG': <bound method Func.from_arg_list of <class 'sqlglot.expressions.LogicalAnd'>>, 'LOGICAL_OR': <bound method Func.from_arg_list of <class 'sqlglot.expressions.LogicalOr'>>, 'BOOL_OR': <bound method Func.from_arg_list of <class 'sqlglot.expressions.LogicalOr'>>, 'BOOLOR_AGG': <bound method Func.from_arg_list of <class 'sqlglot.expressions.LogicalOr'>>, 'LOWER': <function build_lower>, 'LCASE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Lower'>>, 'LOWER_HEX': <bound method Func.from_arg_list of <class 'sqlglot.expressions.LowerHex'>>, 'MD5': <bound method Func.from_arg_list of <class 'sqlglot.expressions.MD5'>>, 'MD5_DIGEST': <bound method Func.from_arg_list of <class 'sqlglot.expressions.MD5Digest'>>, 'MAKE_INTERVAL': <bound method Func.from_arg_list of <class 'sqlglot.expressions.MakeInterval'>>, 'MAP': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Map'>>, 'MAP_FROM_ENTRIES': <bound method Func.from_arg_list of <class 'sqlglot.expressions.MapFromEntries'>>, 'MATCH_AGAINST': <bound method Func.from_arg_list of <class 'sqlglot.expressions.MatchAgainst'>>, 'MAX': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Max'>>, 'MEDIAN': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Median'>>, 'MIN': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Min'>>, 'MONTH': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Month'>>, 'MONTHS_BETWEEN': <bound method Func.from_arg_list of <class 'sqlglot.expressions.MonthsBetween'>>, 'NEXT_VALUE_FOR': <bound method Func.from_arg_list of <class 'sqlglot.expressions.NextValueFor'>>, 'NORMALIZE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Normalize'>>, 'NTH_VALUE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.NthValue'>>, 'NULLIF': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Nullif'>>, 'NUMBER_TO_STR': <bound method Func.from_arg_list of <class 'sqlglot.expressions.NumberToStr'>>, 'NVL2': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Nvl2'>>, 'OBJECT_INSERT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ObjectInsert'>>, 'OPEN_J_S_O_N': <bound method Func.from_arg_list of <class 'sqlglot.expressions.OpenJSON'>>, 'OVERLAY': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Overlay'>>, 'PAD': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Pad'>>, 'PARAMETERIZED_AGG': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ParameterizedAgg'>>, 'PARSE_JSON': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ParseJSON'>>, 'JSON_PARSE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ParseJSON'>>, 'PERCENTILE_CONT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.PercentileCont'>>, 'PERCENTILE_DISC': <bound method Func.from_arg_list of <class 'sqlglot.expressions.PercentileDisc'>>, 'POSEXPLODE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Posexplode'>>, 'POSEXPLODE_OUTER': <bound method Func.from_arg_list of <class 'sqlglot.expressions.PosexplodeOuter'>>, 'POWER': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Pow'>>, 'POW': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Pow'>>, 'PREDICT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Predict'>>, 'QUANTILE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Quantile'>>, 'QUARTER': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Quarter'>>, 'RAND': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Rand'>>, 'RANDOM': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Rand'>>, 'RANDN': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Randn'>>, 'RANGE_N': <bound method Func.from_arg_list of <class 'sqlglot.expressions.RangeN'>>, 'READ_CSV': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ReadCSV'>>, 'REDUCE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Reduce'>>, 'REGEXP_EXTRACT': <function build_regexp_extract.<locals>._builder>, 'REGEXP_EXTRACT_ALL': <function build_regexp_extract.<locals>._builder>, 'REGEXP_I_LIKE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.RegexpILike'>>, 'REGEXP_LIKE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.RegexpLike'>>, 'REGEXP_REPLACE': <function DuckDB.Parser.<lambda>>, 'REGEXP_SPLIT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.RegexpSplit'>>, 'REPEAT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Repeat'>>, 'RIGHT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Right'>>, 'ROUND': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Round'>>, 'ROW_NUMBER': <bound method Func.from_arg_list of <class 'sqlglot.expressions.RowNumber'>>, 'SHA': <bound method Func.from_arg_list of <class 'sqlglot.expressions.SHA'>>, 'SHA1': <bound method Func.from_arg_list of <class 'sqlglot.expressions.SHA'>>, 'SHA2': <bound method Func.from_arg_list of <class 'sqlglot.expressions.SHA2'>>, 'SAFE_DIVIDE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.SafeDivide'>>, 'SIGN': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Sign'>>, 'SIGNUM': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Sign'>>, 'SORT_ARRAY': <bound method Func.from_arg_list of <class 'sqlglot.expressions.SortArray'>>, 'SPLIT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Split'>>, 'SPLIT_PART': <bound method Func.from_arg_list of <class 'sqlglot.expressions.SplitPart'>>, 'SQRT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Sqrt'>>, 'STANDARD_HASH': <bound method Func.from_arg_list of <class 'sqlglot.expressions.StandardHash'>>, 'STAR_MAP': <bound method Func.from_arg_list of <class 'sqlglot.expressions.StarMap'>>, 'STARTS_WITH': <bound method Func.from_arg_list of <class 'sqlglot.expressions.StartsWith'>>, 'STARTSWITH': <bound method Func.from_arg_list of <class 'sqlglot.expressions.StartsWith'>>, 'STDDEV': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Stddev'>>, 'STDEV': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Stddev'>>, 'STDDEV_POP': <bound method Func.from_arg_list of <class 'sqlglot.expressions.StddevPop'>>, 'STDDEV_SAMP': <bound method Func.from_arg_list of <class 'sqlglot.expressions.StddevSamp'>>, 'STR_POSITION': <bound method Func.from_arg_list of <class 'sqlglot.expressions.StrPosition'>>, 'STR_TO_DATE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.StrToDate'>>, 'STR_TO_MAP': <bound method Func.from_arg_list of <class 'sqlglot.expressions.StrToMap'>>, 'STR_TO_TIME': <bound method Func.from_arg_list of <class 'sqlglot.expressions.StrToTime'>>, 'STR_TO_UNIX': <bound method Func.from_arg_list of <class 'sqlglot.expressions.StrToUnix'>>, 'STRING': <bound method Func.from_arg_list of <class 'sqlglot.expressions.String'>>, 'STRING_TO_ARRAY': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Split'>>, 'SPLIT_BY_STRING': <bound method Func.from_arg_list of <class 'sqlglot.expressions.StringToArray'>>, 'STRUCT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Struct'>>, 'STRUCT_EXTRACT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.StructExtract'>>, 'STUFF': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Stuff'>>, 'INSERT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Stuff'>>, 'SUBSTRING': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Substring'>>, 'SUBSTR': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Substring'>>, 'SUM': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Sum'>>, 'TIME': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Time'>>, 'TIME_ADD': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TimeAdd'>>, 'TIME_DIFF': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TimeDiff'>>, 'TIME_FROM_PARTS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TimeFromParts'>>, 'TIMEFROMPARTS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TimeFromParts'>>, 'TIME_STR_TO_DATE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TimeStrToDate'>>, 'TIME_STR_TO_TIME': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TimeStrToTime'>>, 'TIME_STR_TO_UNIX': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TimeStrToUnix'>>, 'TIME_SUB': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TimeSub'>>, 'TIME_TO_STR': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TimeToStr'>>, 'TIME_TO_TIME_STR': <function Parser.<lambda>>, 'TIME_TO_UNIX': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TimeToUnix'>>, 'TIME_TRUNC': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TimeTrunc'>>, 'TIMESTAMP': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Timestamp'>>, 'TIMESTAMP_ADD': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TimestampAdd'>>, 'TIMESTAMPDIFF': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TimestampDiff'>>, 'TIMESTAMP_DIFF': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TimestampDiff'>>, 'TIMESTAMP_FROM_PARTS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TimestampFromParts'>>, 'TIMESTAMPFROMPARTS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TimestampFromParts'>>, 'TIMESTAMP_SUB': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TimestampSub'>>, 'TIMESTAMP_TRUNC': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TimestampTrunc'>>, 'TO_ARRAY': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ToArray'>>, 'TO_BASE64': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ToBase64'>>, 'TO_CHAR': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ToChar'>>, 'TO_DAYS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ToDays'>>, 'TO_DOUBLE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ToDouble'>>, 'TO_MAP': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ToMap'>>, 'TO_NUMBER': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ToNumber'>>, 'TRANSFORM': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Transform'>>, 'TRIM': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Trim'>>, 'TRY': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Try'>>, 'TRY_CAST': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TryCast'>>, 'TS_OR_DI_TO_DI': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TsOrDiToDi'>>, 'TS_OR_DS_ADD': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TsOrDsAdd'>>, 'TS_OR_DS_DIFF': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TsOrDsDiff'>>, 'TS_OR_DS_TO_DATE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TsOrDsToDate'>>, 'TS_OR_DS_TO_DATE_STR': <function Parser.<lambda>>, 'TS_OR_DS_TO_DATETIME': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TsOrDsToDatetime'>>, 'TS_OR_DS_TO_TIME': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TsOrDsToTime'>>, 'TS_OR_DS_TO_TIMESTAMP': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TsOrDsToTimestamp'>>, 'UNHEX': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Unhex'>>, 'UNICODE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Unicode'>>, 'UNIX_DATE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.UnixDate'>>, 'UNIX_SECONDS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.UnixSeconds'>>, 'UNIX_TO_STR': <bound method Func.from_arg_list of <class 'sqlglot.expressions.UnixToStr'>>, 'UNIX_TO_TIME': <bound method Func.from_arg_list of <class 'sqlglot.expressions.UnixToTime'>>, 'UNIX_TO_TIME_STR': <bound method Func.from_arg_list of <class 'sqlglot.expressions.UnixToTimeStr'>>, 'UNNEST': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Explode'>>, 'UPPER': <function build_upper>, 'UCASE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Upper'>>, 'UUID': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Uuid'>>, 'GEN_RANDOM_UUID': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Uuid'>>, 'GENERATE_UUID': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Uuid'>>, 'UUID_STRING': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Uuid'>>, 'VAR_MAP': <function build_var_map>, 'VARIANCE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Variance'>>, 'VARIANCE_SAMP': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Variance'>>, 'VAR_SAMP': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Variance'>>, 'VARIANCE_POP': <bound method Func.from_arg_list of <class 'sqlglot.expressions.VariancePop'>>, 'VAR_POP': <bound method Func.from_arg_list of <class 'sqlglot.expressions.VariancePop'>>, 'WEEK': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Week'>>, 'WEEK_OF_YEAR': <bound method Func.from_arg_list of <class 'sqlglot.expressions.WeekOfYear'>>, 'WEEKOFYEAR': <bound method Func.from_arg_list of <class 'sqlglot.expressions.WeekOfYear'>>, 'XMLELEMENT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.XMLElement'>>, 'X_M_L_TABLE': <bound method Func.from_arg_list of <class 'sqlglot.expressions.XMLTable'>>, 'XOR': <function binary_from_function.<locals>.<lambda>>, 'YEAR': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Year'>>, 'ARRAYAGG': <function Parser.<lambda>>, 'JSON_EXTRACT_PATH_TEXT': <function build_extract_json_with_path.<locals>._builder>, 'LIKE': <function build_like>, 'LOG2': <function Parser.<lambda>>, 'LOG10': <function Parser.<lambda>>, 'LPAD': <function Parser.<lambda>>, 'LEFTPAD': <function Parser.<lambda>>, 'LTRIM': <function Parser.<lambda>>, 'MOD': <function build_mod>, 'RIGHTPAD': <function Parser.<lambda>>, 'RPAD': <function Parser.<lambda>>, 'RTRIM': <function Parser.<lambda>>, 'SCOPE_RESOLUTION': <function Parser.<lambda>>, 'STRPOS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.StrPosition'>>, 'CHARINDEX': <function Parser.<lambda>>, 'INSTR': <bound method Func.from_arg_list of <class 'sqlglot.expressions.StrPosition'>>, 'LOCATE': <function Parser.<lambda>>, 'TO_HEX': <function build_hex>, 'ARRAY_REVERSE_SORT': <function _build_sort_array_desc>, 'DATETRUNC': <function date_trunc_to_time>, 'EPOCH': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TimeToUnix'>>, 'EPOCH_MS': <function DuckDB.Parser.<lambda>>, 'JSON': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ParseJSON'>>, 'JSON_EXTRACT_PATH': <function build_extract_json_with_path.<locals>._builder>, 'JSON_EXTRACT_STRING': <function build_extract_json_with_path.<locals>._builder>, 'LIST_HAS': <bound method Func.from_arg_list of <class 'sqlglot.expressions.ArrayContains'>>, 'LIST_REVERSE_SORT': <function _build_sort_array_desc>, 'LIST_SORT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.SortArray'>>, 'LIST_VALUE': <function DuckDB.Parser.<lambda>>, 'MAKE_TIME': <bound method Func.from_arg_list of <class 'sqlglot.expressions.TimeFromParts'>>, 'MAKE_TIMESTAMP': <function _build_make_timestamp>, 'QUANTILE_CONT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.PercentileCont'>>, 'QUANTILE_DISC': <bound method Func.from_arg_list of <class 'sqlglot.expressions.PercentileDisc'>>, 'REGEXP_MATCHES': <bound method Func.from_arg_list of <class 'sqlglot.expressions.RegexpLike'>>, 'STRFTIME': <function build_formatted_time.<locals>._builder>, 'STRING_SPLIT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Split'>>, 'STRING_SPLIT_REGEX': <bound method Func.from_arg_list of <class 'sqlglot.expressions.RegexpSplit'>>, 'STRPTIME': <function build_formatted_time.<locals>._builder>, 'STRUCT_PACK': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Struct'>>, 'STR_SPLIT': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Split'>>, 'STR_SPLIT_REGEX': <bound method Func.from_arg_list of <class 'sqlglot.expressions.RegexpSplit'>>, 'TO_TIMESTAMP': <bound method Func.from_arg_list of <class 'sqlglot.expressions.UnixToTime'>>, 'RANGE': <function _build_generate_series.<locals>._builder>, 'EDITDIST3': <bound method Func.from_arg_list of <class 'sqlglot.expressions.Levenshtein'>>}
FUNCTION_PARSERS = {'CAST': <function Parser.<lambda>>, 'CEIL': <function Parser.<lambda>>, 'CONVERT': <function Parser.<lambda>>, 'EXTRACT': <function Parser.<lambda>>, 'FLOOR': <function Parser.<lambda>>, 'GAP_FILL': <function Parser.<lambda>>, 'JSON_OBJECT': <function Parser.<lambda>>, 'JSON_OBJECTAGG': <function Parser.<lambda>>, 'JSON_TABLE': <function Parser.<lambda>>, 'MATCH': <function Parser.<lambda>>, 'NORMALIZE': <function Parser.<lambda>>, 'OPENJSON': <function Parser.<lambda>>, 'OVERLAY': <function Parser.<lambda>>, 'POSITION': <function Parser.<lambda>>, 'PREDICT': <function Parser.<lambda>>, 'SAFE_CAST': <function Parser.<lambda>>, 'STRING_AGG': <function Parser.<lambda>>, 'SUBSTRING': <function Parser.<lambda>>, 'TRIM': <function Parser.<lambda>>, 'TRY_CAST': <function Parser.<lambda>>, 'TRY_CONVERT': <function Parser.<lambda>>, 'XMLELEMENT': <function Parser.<lambda>>, 'XMLTABLE': <function Parser.<lambda>>}
NO_PAREN_FUNCTION_PARSERS = {'ANY': <function Parser.<lambda>>, 'CASE': <function Parser.<lambda>>, 'CONNECT_BY_ROOT': <function Parser.<lambda>>, 'IF': <function Parser.<lambda>>, 'MAP': <function DuckDB.Parser.<lambda>>}
TABLE_ALIAS_TOKENS = {<TokenType.NEXT: 'NEXT'>, <TokenType.UNNEST: 'UNNEST'>, <TokenType.STORAGE_INTEGRATION: 'STORAGE_INTEGRATION'>, <TokenType.NVARCHAR: 'NVARCHAR'>, <TokenType.REFERENCES: 'REFERENCES'>, <TokenType.ROW: 'ROW'>, <TokenType.TSTZMULTIRANGE: 'TSTZMULTIRANGE'>, <TokenType.POINT: 'POINT'>, <TokenType.ENUM8: 'ENUM8'>, <TokenType.BIGDECIMAL: 'BIGDECIMAL'>, <TokenType.CASE: 'CASE'>, <TokenType.TIME: 'TIME'>, <TokenType.LOAD: 'LOAD'>, <TokenType.OFFSET: 'OFFSET'>, <TokenType.TIMESTAMP: 'TIMESTAMP'>, <TokenType.DOUBLE: 'DOUBLE'>, <TokenType.JSONB: 'JSONB'>, <TokenType.PIVOT: 'PIVOT'>, <TokenType.UMEDIUMINT: 'UMEDIUMINT'>, <TokenType.SEQUENCE: 'SEQUENCE'>, <TokenType.MERGE: 'MERGE'>, <TokenType.SOURCE: 'SOURCE'>, <TokenType.GEOGRAPHY: 'GEOGRAPHY'>, <TokenType.ASC: 'ASC'>, <TokenType.CONSTRAINT: 'CONSTRAINT'>, <TokenType.MODEL: 'MODEL'>, <TokenType.TIMESTAMP_NS: 'TIMESTAMP_NS'>, <TokenType.PRAGMA: 'PRAGMA'>, <TokenType.SMALLMONEY: 'SMALLMONEY'>, <TokenType.VOLATILE: 'VOLATILE'>, <TokenType.NULLABLE: 'NULLABLE'>, <TokenType.OVERLAPS: 'OVERLAPS'>, <TokenType.INTERVAL: 'INTERVAL'>, <TokenType.DATEMULTIRANGE: 'DATEMULTIRANGE'>, <TokenType.FOREIGN_KEY: 'FOREIGN_KEY'>, <TokenType.SET: 'SET'>, <TokenType.VIEW: 'VIEW'>, <TokenType.INT: 'INT'>, <TokenType.ENUM16: 'ENUM16'>, <TokenType.JSON: 'JSON'>, <TokenType.SMALLINT: 'SMALLINT'>, <TokenType.EXECUTE: 'EXECUTE'>, <TokenType.NUMRANGE: 'NUMRANGE'>, <TokenType.DATETIME: 'DATETIME'>, <TokenType.COPY: 'COPY'>, <TokenType.BINARY: 'BINARY'>, <TokenType.SCHEMA: 'SCHEMA'>, <TokenType.TINYINT: 'TINYINT'>, <TokenType.COMMENT: 'COMMENT'>, <TokenType.INT128: 'INT128'>, <TokenType.UDECIMAL: 'UDECIMAL'>, <TokenType.TOP: 'TOP'>, <TokenType.ESCAPE: 'ESCAPE'>, <TokenType.TIMESTAMPNTZ: 'TIMESTAMPNTZ'>, <TokenType.DELETE: 'DELETE'>, <TokenType.DICTIONARY: 'DICTIONARY'>, <TokenType.COLUMN: 'COLUMN'>, <TokenType.LINESTRING: 'LINESTRING'>, <TokenType.UINT256: 'UINT256'>, <TokenType.ROWVERSION: 'ROWVERSION'>, <TokenType.CURRENT_USER: 'CURRENT_USER'>, <TokenType.EXISTS: 'EXISTS'>, <TokenType.PSEUDO_TYPE: 'PSEUDO_TYPE'>, <TokenType.UNPIVOT: 'UNPIVOT'>, <TokenType.RENAME: 'RENAME'>, <TokenType.REFRESH: 'REFRESH'>, <TokenType.NAME: 'NAME'>, <TokenType.LONGBLOB: 'LONGBLOB'>, <TokenType.MONEY: 'MONEY'>, <TokenType.DATE32: 'DATE32'>, <TokenType.TIMESTAMP_S: 'TIMESTAMP_S'>, <TokenType.HSTORE: 'HSTORE'>, <TokenType.ANY: 'ANY'>, <TokenType.RING: 'RING'>, <TokenType.PERCENT: 'PERCENT'>, <TokenType.DATE: 'DATE'>, <TokenType.NULL: 'NULL'>, <TokenType.FIRST: 'FIRST'>, <TokenType.KILL: 'KILL'>, <TokenType.TEMPORARY: 'TEMPORARY'>, <TokenType.UNIQUEIDENTIFIER: 'UNIQUEIDENTIFIER'>, <TokenType.LIMIT: 'LIMIT'>, <TokenType.TIMETZ: 'TIMETZ'>, <TokenType.DIV: 'DIV'>, <TokenType.OBJECT: 'OBJECT'>, <TokenType.CUBE: 'CUBE'>, <TokenType.DEFAULT: 'DEFAULT'>, <TokenType.DECIMAL128: 'DECIMAL128'>, <TokenType.UINT: 'UINT'>, <TokenType.TSMULTIRANGE: 'TSMULTIRANGE'>, <TokenType.VAR: 'VAR'>, <TokenType.OVERWRITE: 'OVERWRITE'>, <TokenType.UTINYINT: 'UTINYINT'>, <TokenType.SOME: 'SOME'>, <TokenType.DESCRIBE: 'DESCRIBE'>, <TokenType.TIMESTAMP_MS: 'TIMESTAMP_MS'>, <TokenType.IPV4: 'IPV4'>, <TokenType.TIMESTAMPLTZ: 'TIMESTAMPLTZ'>, <TokenType.INET: 'INET'>, <TokenType.TDIGEST: 'TDIGEST'>, <TokenType.DATETIME2: 'DATETIME2'>, <TokenType.FALSE: 'FALSE'>, <TokenType.OBJECT_IDENTIFIER: 'OBJECT_IDENTIFIER'>, <TokenType.DETACH: 'DETACH'>, <TokenType.YEAR: 'YEAR'>, <TokenType.IMAGE: 'IMAGE'>, <TokenType.FLOAT: 'FLOAT'>, <TokenType.TSTZRANGE: 'TSTZRANGE'>, <TokenType.VARCHAR: 'VARCHAR'>, <TokenType.MEDIUMBLOB: 'MEDIUMBLOB'>, <TokenType.CURRENT_TIMESTAMP: 'CURRENT_TIMESTAMP'>, <TokenType.DATABASE: 'DATABASE'>, <TokenType.LOWCARDINALITY: 'LOWCARDINALITY'>, <TokenType.RECURSIVE: 'RECURSIVE'>, <TokenType.TABLE: 'TABLE'>, <TokenType.CURRENT_DATE: 'CURRENT_DATE'>, <TokenType.STRUCT: 'STRUCT'>, <TokenType.COLLATE: 'COLLATE'>, <TokenType.DATETIME64: 'DATETIME64'>, <TokenType.ROLLUP: 'ROLLUP'>, <TokenType.INT4RANGE: 'INT4RANGE'>, <TokenType.COMMAND: 'COMMAND'>, <TokenType.IDENTIFIER: 'IDENTIFIER'>, <TokenType.UINT128: 'UINT128'>, <TokenType.TEXT: 'TEXT'>, <TokenType.KEEP: 'KEEP'>, <TokenType.VARBINARY: 'VARBINARY'>, <TokenType.ARRAY: 'ARRAY'>, <TokenType.FORMAT: 'FORMAT'>, <TokenType.UNKNOWN: 'UNKNOWN'>, <TokenType.DATERANGE: 'DATERANGE'>, <TokenType.COMMIT: 'COMMIT'>, <TokenType.ATTACH: 'ATTACH'>, <TokenType.SERIAL: 'SERIAL'>, <TokenType.INT8RANGE: 'INT8RANGE'>, <TokenType.TRUE: 'TRUE'>, <TokenType.ENUM: 'ENUM'>, <TokenType.PARTITION: 'PARTITION'>, <TokenType.WAREHOUSE: 'WAREHOUSE'>, <TokenType.CHAR: 'CHAR'>, <TokenType.UPDATE: 'UPDATE'>, <TokenType.IPADDRESS: 'IPADDRESS'>, <TokenType.INDEX: 'INDEX'>, <TokenType.AUTO_INCREMENT: 'AUTO_INCREMENT'>, <TokenType.PROCEDURE: 'PROCEDURE'>, <TokenType.REPLACE: 'REPLACE'>, <TokenType.USE: 'USE'>, <TokenType.TAG: 'TAG'>, <TokenType.CACHE: 'CACHE'>, <TokenType.FUNCTION: 'FUNCTION'>, <TokenType.BIGSERIAL: 'BIGSERIAL'>, <TokenType.MULTIPOLYGON: 'MULTIPOLYGON'>, <TokenType.INT8MULTIRANGE: 'INT8MULTIRANGE'>, <TokenType.HLLSKETCH: 'HLLSKETCH'>, <TokenType.MULTILINESTRING: 'MULTILINESTRING'>, <TokenType.OPERATOR: 'OPERATOR'>, <TokenType.DECIMAL: 'DECIMAL'>, <TokenType.TINYTEXT: 'TINYTEXT'>, <TokenType.VARIANT: 'VARIANT'>, <TokenType.UNIQUE: 'UNIQUE'>, <TokenType.USERDEFINED: 'USERDEFINED'>, <TokenType.MEDIUMTEXT: 'MEDIUMTEXT'>, <TokenType.SHOW: 'SHOW'>, <TokenType.TIMESTAMPTZ: 'TIMESTAMPTZ'>, <TokenType.STREAMLIT: 'STREAMLIT'>, <TokenType.SMALLDATETIME: 'SMALLDATETIME'>, <TokenType.DECIMAL256: 'DECIMAL256'>, <TokenType.LIST: 'LIST'>, <TokenType.TSRANGE: 'TSRANGE'>, <TokenType.DECIMAL64: 'DECIMAL64'>, <TokenType.NCHAR: 'NCHAR'>, <TokenType.TINYBLOB: 'TINYBLOB'>, <TokenType.BIGINT: 'BIGINT'>, <TokenType.ORDINALITY: 'ORDINALITY'>, <TokenType.NAMESPACE: 'NAMESPACE'>, <TokenType.IPV6: 'IPV6'>, <TokenType.IPPREFIX: 'IPPREFIX'>, <TokenType.BPCHAR: 'BPCHAR'>, <TokenType.SMALLSERIAL: 'SMALLSERIAL'>, <TokenType.SIMPLEAGGREGATEFUNCTION: 'SIMPLEAGGREGATEFUNCTION'>, <TokenType.DESC: 'DESC'>, <TokenType.GEOMETRY: 'GEOMETRY'>, <TokenType.STRAIGHT_JOIN: 'STRAIGHT_JOIN'>, <TokenType.END: 'END'>, <TokenType.XML: 'XML'>, <TokenType.ISNULL: 'ISNULL'>, <TokenType.FILTER: 'FILTER'>, <TokenType.USMALLINT: 'USMALLINT'>, <TokenType.IS: 'IS'>, <TokenType.DECIMAL32: 'DECIMAL32'>, <TokenType.FIXEDSTRING: 'FIXEDSTRING'>, <TokenType.BOOLEAN: 'BOOLEAN'>, <TokenType.DYNAMIC: 'DYNAMIC'>, <TokenType.POLYGON: 'POLYGON'>, <TokenType.NESTED: 'NESTED'>, <TokenType.SETTINGS: 'SETTINGS'>, <TokenType.BIT: 'BIT'>, <TokenType.INT256: 'INT256'>, <TokenType.MEDIUMINT: 'MEDIUMINT'>, <TokenType.BEGIN: 'BEGIN'>, <TokenType.LONGTEXT: 'LONGTEXT'>, <TokenType.FINAL: 'FINAL'>, <TokenType.UBIGINT: 'UBIGINT'>, <TokenType.CURRENT_DATETIME: 'CURRENT_DATETIME'>, <TokenType.SUPER: 'SUPER'>, <TokenType.SINK: 'SINK'>, <TokenType.CURRENT_TIME: 'CURRENT_TIME'>, <TokenType.RANGE: 'RANGE'>, <TokenType.ALL: 'ALL'>, <TokenType.VECTOR: 'VECTOR'>, <TokenType.TRUNCATE: 'TRUNCATE'>, <TokenType.NUMMULTIRANGE: 'NUMMULTIRANGE'>, <TokenType.AGGREGATEFUNCTION: 'AGGREGATEFUNCTION'>, <TokenType.INT4MULTIRANGE: 'INT4MULTIRANGE'>, <TokenType.UUID: 'UUID'>, <TokenType.ROWS: 'ROWS'>, <TokenType.MAP: 'MAP'>}
PLACEHOLDER_PARSERS = {<TokenType.PLACEHOLDER: 'PLACEHOLDER'>: <function Parser.<lambda>>, <TokenType.PARAMETER: 'PARAMETER'>: <function DuckDB.Parser.<lambda>>, <TokenType.COLON: 'COLON'>: <function Parser.<lambda>>}
TYPE_CONVERTERS = {<Type.DECIMAL: 'DECIMAL'>: <function build_default_decimal_type.<locals>._builder>, <Type.TEXT: 'TEXT'>: <function DuckDB.Parser.<lambda>>}
STATEMENT_PARSERS = {<TokenType.ALTER: 'ALTER'>: <function Parser.<lambda>>, <TokenType.ANALYZE: 'ANALYZE'>: <function Parser.<lambda>>, <TokenType.BEGIN: 'BEGIN'>: <function Parser.<lambda>>, <TokenType.CACHE: 'CACHE'>: <function Parser.<lambda>>, <TokenType.COMMENT: 'COMMENT'>: <function Parser.<lambda>>, <TokenType.COMMIT: 'COMMIT'>: <function Parser.<lambda>>, <TokenType.COPY: 'COPY'>: <function Parser.<lambda>>, <TokenType.CREATE: 'CREATE'>: <function Parser.<lambda>>, <TokenType.DELETE: 'DELETE'>: <function Parser.<lambda>>, <TokenType.DESC: 'DESC'>: <function Parser.<lambda>>, <TokenType.DESCRIBE: 'DESCRIBE'>: <function Parser.<lambda>>, <TokenType.DROP: 'DROP'>: <function Parser.<lambda>>, <TokenType.GRANT: 'GRANT'>: <function Parser.<lambda>>, <TokenType.INSERT: 'INSERT'>: <function Parser.<lambda>>, <TokenType.KILL: 'KILL'>: <function Parser.<lambda>>, <TokenType.LOAD: 'LOAD'>: <function Parser.<lambda>>, <TokenType.MERGE: 'MERGE'>: <function Parser.<lambda>>, <TokenType.PIVOT: 'PIVOT'>: <function Parser.<lambda>>, <TokenType.PRAGMA: 'PRAGMA'>: <function Parser.<lambda>>, <TokenType.REFRESH: 'REFRESH'>: <function Parser.<lambda>>, <TokenType.ROLLBACK: 'ROLLBACK'>: <function Parser.<lambda>>, <TokenType.SET: 'SET'>: <function Parser.<lambda>>, <TokenType.TRUNCATE: 'TRUNCATE'>: <function Parser.<lambda>>, <TokenType.UNCACHE: 'UNCACHE'>: <function Parser.<lambda>>, <TokenType.UNPIVOT: 'UNPIVOT'>: <function Parser.<lambda>>, <TokenType.UPDATE: 'UPDATE'>: <function Parser.<lambda>>, <TokenType.USE: 'USE'>: <function Parser.<lambda>>, <TokenType.SEMICOLON: 'SEMICOLON'>: <function Parser.<lambda>>, <TokenType.ATTACH: 'ATTACH'>: <function DuckDB.Parser.<lambda>>, <TokenType.DETACH: 'DETACH'>: <function DuckDB.Parser.<lambda>>}
ID_VAR_TOKENS = {<TokenType.NEXT: 'NEXT'>, <TokenType.UNNEST: 'UNNEST'>, <TokenType.STORAGE_INTEGRATION: 'STORAGE_INTEGRATION'>, <TokenType.NVARCHAR: 'NVARCHAR'>, <TokenType.REFERENCES: 'REFERENCES'>, <TokenType.ROW: 'ROW'>, <TokenType.TSTZMULTIRANGE: 'TSTZMULTIRANGE'>, <TokenType.POINT: 'POINT'>, <TokenType.ENUM8: 'ENUM8'>, <TokenType.BIGDECIMAL: 'BIGDECIMAL'>, <TokenType.CASE: 'CASE'>, <TokenType.TIME: 'TIME'>, <TokenType.LOAD: 'LOAD'>, <TokenType.OFFSET: 'OFFSET'>, <TokenType.TIMESTAMP: 'TIMESTAMP'>, <TokenType.DOUBLE: 'DOUBLE'>, <TokenType.JSONB: 'JSONB'>, <TokenType.PIVOT: 'PIVOT'>, <TokenType.UMEDIUMINT: 'UMEDIUMINT'>, <TokenType.SEQUENCE: 'SEQUENCE'>, <TokenType.MERGE: 'MERGE'>, <TokenType.SOURCE: 'SOURCE'>, <TokenType.GEOGRAPHY: 'GEOGRAPHY'>, <TokenType.ASC: 'ASC'>, <TokenType.CONSTRAINT: 'CONSTRAINT'>, <TokenType.MODEL: 'MODEL'>, <TokenType.TIMESTAMP_NS: 'TIMESTAMP_NS'>, <TokenType.PRAGMA: 'PRAGMA'>, <TokenType.SMALLMONEY: 'SMALLMONEY'>, <TokenType.VOLATILE: 'VOLATILE'>, <TokenType.NULLABLE: 'NULLABLE'>, <TokenType.OVERLAPS: 'OVERLAPS'>, <TokenType.INTERVAL: 'INTERVAL'>, <TokenType.DATEMULTIRANGE: 'DATEMULTIRANGE'>, <TokenType.FOREIGN_KEY: 'FOREIGN_KEY'>, <TokenType.SET: 'SET'>, <TokenType.VIEW: 'VIEW'>, <TokenType.INT: 'INT'>, <TokenType.ENUM16: 'ENUM16'>, <TokenType.FULL: 'FULL'>, <TokenType.JSON: 'JSON'>, <TokenType.SMALLINT: 'SMALLINT'>, <TokenType.EXECUTE: 'EXECUTE'>, <TokenType.NUMRANGE: 'NUMRANGE'>, <TokenType.DATETIME: 'DATETIME'>, <TokenType.COPY: 'COPY'>, <TokenType.BINARY: 'BINARY'>, <TokenType.SCHEMA: 'SCHEMA'>, <TokenType.TINYINT: 'TINYINT'>, <TokenType.COMMENT: 'COMMENT'>, <TokenType.INT128: 'INT128'>, <TokenType.UDECIMAL: 'UDECIMAL'>, <TokenType.TOP: 'TOP'>, <TokenType.ESCAPE: 'ESCAPE'>, <TokenType.TIMESTAMPNTZ: 'TIMESTAMPNTZ'>, <TokenType.DELETE: 'DELETE'>, <TokenType.DICTIONARY: 'DICTIONARY'>, <TokenType.COLUMN: 'COLUMN'>, <TokenType.LINESTRING: 'LINESTRING'>, <TokenType.UINT256: 'UINT256'>, <TokenType.ROWVERSION: 'ROWVERSION'>, <TokenType.CURRENT_USER: 'CURRENT_USER'>, <TokenType.EXISTS: 'EXISTS'>, <TokenType.PSEUDO_TYPE: 'PSEUDO_TYPE'>, <TokenType.UNPIVOT: 'UNPIVOT'>, <TokenType.RENAME: 'RENAME'>, <TokenType.REFRESH: 'REFRESH'>, <TokenType.NAME: 'NAME'>, <TokenType.LONGBLOB: 'LONGBLOB'>, <TokenType.MONEY: 'MONEY'>, <TokenType.DATE32: 'DATE32'>, <TokenType.TIMESTAMP_S: 'TIMESTAMP_S'>, <TokenType.HSTORE: 'HSTORE'>, <TokenType.ANY: 'ANY'>, <TokenType.RING: 'RING'>, <TokenType.PERCENT: 'PERCENT'>, <TokenType.NATURAL: 'NATURAL'>, <TokenType.DATE: 'DATE'>, <TokenType.NULL: 'NULL'>, <TokenType.LEFT: 'LEFT'>, <TokenType.FIRST: 'FIRST'>, <TokenType.KILL: 'KILL'>, <TokenType.TEMPORARY: 'TEMPORARY'>, <TokenType.UNIQUEIDENTIFIER: 'UNIQUEIDENTIFIER'>, <TokenType.LIMIT: 'LIMIT'>, <TokenType.TIMETZ: 'TIMETZ'>, <TokenType.DIV: 'DIV'>, <TokenType.OBJECT: 'OBJECT'>, <TokenType.CUBE: 'CUBE'>, <TokenType.DEFAULT: 'DEFAULT'>, <TokenType.DECIMAL128: 'DECIMAL128'>, <TokenType.UINT: 'UINT'>, <TokenType.TSMULTIRANGE: 'TSMULTIRANGE'>, <TokenType.ROWS: 'ROWS'>, <TokenType.VAR: 'VAR'>, <TokenType.OVERWRITE: 'OVERWRITE'>, <TokenType.UTINYINT: 'UTINYINT'>, <TokenType.SOME: 'SOME'>, <TokenType.ANTI: 'ANTI'>, <TokenType.DESCRIBE: 'DESCRIBE'>, <TokenType.TIMESTAMP_MS: 'TIMESTAMP_MS'>, <TokenType.IPV4: 'IPV4'>, <TokenType.TIMESTAMPLTZ: 'TIMESTAMPLTZ'>, <TokenType.INET: 'INET'>, <TokenType.TDIGEST: 'TDIGEST'>, <TokenType.DATETIME2: 'DATETIME2'>, <TokenType.FALSE: 'FALSE'>, <TokenType.OBJECT_IDENTIFIER: 'OBJECT_IDENTIFIER'>, <TokenType.DETACH: 'DETACH'>, <TokenType.YEAR: 'YEAR'>, <TokenType.IMAGE: 'IMAGE'>, <TokenType.FLOAT: 'FLOAT'>, <TokenType.TSTZRANGE: 'TSTZRANGE'>, <TokenType.VARCHAR: 'VARCHAR'>, <TokenType.MEDIUMBLOB: 'MEDIUMBLOB'>, <TokenType.CURRENT_TIMESTAMP: 'CURRENT_TIMESTAMP'>, <TokenType.DATABASE: 'DATABASE'>, <TokenType.WINDOW: 'WINDOW'>, <TokenType.LOWCARDINALITY: 'LOWCARDINALITY'>, <TokenType.RECURSIVE: 'RECURSIVE'>, <TokenType.TABLE: 'TABLE'>, <TokenType.CURRENT_DATE: 'CURRENT_DATE'>, <TokenType.STRUCT: 'STRUCT'>, <TokenType.COLLATE: 'COLLATE'>, <TokenType.DATETIME64: 'DATETIME64'>, <TokenType.ROLLUP: 'ROLLUP'>, <TokenType.INT4RANGE: 'INT4RANGE'>, <TokenType.COMMAND: 'COMMAND'>, <TokenType.IDENTIFIER: 'IDENTIFIER'>, <TokenType.UINT128: 'UINT128'>, <TokenType.TEXT: 'TEXT'>, <TokenType.KEEP: 'KEEP'>, <TokenType.VARBINARY: 'VARBINARY'>, <TokenType.ARRAY: 'ARRAY'>, <TokenType.FORMAT: 'FORMAT'>, <TokenType.UNKNOWN: 'UNKNOWN'>, <TokenType.DATERANGE: 'DATERANGE'>, <TokenType.COMMIT: 'COMMIT'>, <TokenType.ATTACH: 'ATTACH'>, <TokenType.SERIAL: 'SERIAL'>, <TokenType.INT8RANGE: 'INT8RANGE'>, <TokenType.TRUE: 'TRUE'>, <TokenType.ENUM: 'ENUM'>, <TokenType.PARTITION: 'PARTITION'>, <TokenType.ASOF: 'ASOF'>, <TokenType.WAREHOUSE: 'WAREHOUSE'>, <TokenType.CHAR: 'CHAR'>, <TokenType.UPDATE: 'UPDATE'>, <TokenType.IPADDRESS: 'IPADDRESS'>, <TokenType.INDEX: 'INDEX'>, <TokenType.AUTO_INCREMENT: 'AUTO_INCREMENT'>, <TokenType.PROCEDURE: 'PROCEDURE'>, <TokenType.REPLACE: 'REPLACE'>, <TokenType.USE: 'USE'>, <TokenType.TAG: 'TAG'>, <TokenType.CACHE: 'CACHE'>, <TokenType.FUNCTION: 'FUNCTION'>, <TokenType.BIGSERIAL: 'BIGSERIAL'>, <TokenType.MULTIPOLYGON: 'MULTIPOLYGON'>, <TokenType.INT8MULTIRANGE: 'INT8MULTIRANGE'>, <TokenType.HLLSKETCH: 'HLLSKETCH'>, <TokenType.MULTILINESTRING: 'MULTILINESTRING'>, <TokenType.OPERATOR: 'OPERATOR'>, <TokenType.DECIMAL: 'DECIMAL'>, <TokenType.TINYTEXT: 'TINYTEXT'>, <TokenType.VARIANT: 'VARIANT'>, <TokenType.UNIQUE: 'UNIQUE'>, <TokenType.USERDEFINED: 'USERDEFINED'>, <TokenType.MEDIUMTEXT: 'MEDIUMTEXT'>, <TokenType.SHOW: 'SHOW'>, <TokenType.TIMESTAMPTZ: 'TIMESTAMPTZ'>, <TokenType.STREAMLIT: 'STREAMLIT'>, <TokenType.SMALLDATETIME: 'SMALLDATETIME'>, <TokenType.DECIMAL256: 'DECIMAL256'>, <TokenType.LIST: 'LIST'>, <TokenType.TSRANGE: 'TSRANGE'>, <TokenType.DECIMAL64: 'DECIMAL64'>, <TokenType.NCHAR: 'NCHAR'>, <TokenType.APPLY: 'APPLY'>, <TokenType.TINYBLOB: 'TINYBLOB'>, <TokenType.BIGINT: 'BIGINT'>, <TokenType.ORDINALITY: 'ORDINALITY'>, <TokenType.NAMESPACE: 'NAMESPACE'>, <TokenType.IPV6: 'IPV6'>, <TokenType.IPPREFIX: 'IPPREFIX'>, <TokenType.BPCHAR: 'BPCHAR'>, <TokenType.SMALLSERIAL: 'SMALLSERIAL'>, <TokenType.SIMPLEAGGREGATEFUNCTION: 'SIMPLEAGGREGATEFUNCTION'>, <TokenType.DESC: 'DESC'>, <TokenType.GEOMETRY: 'GEOMETRY'>, <TokenType.STRAIGHT_JOIN: 'STRAIGHT_JOIN'>, <TokenType.END: 'END'>, <TokenType.XML: 'XML'>, <TokenType.ISNULL: 'ISNULL'>, <TokenType.FILTER: 'FILTER'>, <TokenType.SEMI: 'SEMI'>, <TokenType.IS: 'IS'>, <TokenType.USMALLINT: 'USMALLINT'>, <TokenType.DECIMAL32: 'DECIMAL32'>, <TokenType.FIXEDSTRING: 'FIXEDSTRING'>, <TokenType.BOOLEAN: 'BOOLEAN'>, <TokenType.DYNAMIC: 'DYNAMIC'>, <TokenType.POLYGON: 'POLYGON'>, <TokenType.NESTED: 'NESTED'>, <TokenType.SETTINGS: 'SETTINGS'>, <TokenType.BIT: 'BIT'>, <TokenType.INT256: 'INT256'>, <TokenType.MEDIUMINT: 'MEDIUMINT'>, <TokenType.BEGIN: 'BEGIN'>, <TokenType.LONGTEXT: 'LONGTEXT'>, <TokenType.FINAL: 'FINAL'>, <TokenType.UBIGINT: 'UBIGINT'>, <TokenType.CURRENT_DATETIME: 'CURRENT_DATETIME'>, <TokenType.SUPER: 'SUPER'>, <TokenType.SINK: 'SINK'>, <TokenType.CURRENT_TIME: 'CURRENT_TIME'>, <TokenType.RANGE: 'RANGE'>, <TokenType.ALL: 'ALL'>, <TokenType.VECTOR: 'VECTOR'>, <TokenType.TRUNCATE: 'TRUNCATE'>, <TokenType.NUMMULTIRANGE: 'NUMMULTIRANGE'>, <TokenType.AGGREGATEFUNCTION: 'AGGREGATEFUNCTION'>, <TokenType.INT4MULTIRANGE: 'INT4MULTIRANGE'>, <TokenType.UUID: 'UUID'>, <TokenType.RIGHT: 'RIGHT'>, <TokenType.MAP: 'MAP'>}
SHOW_TRIE: Dict = {}
SET_TRIE: Dict = {'GLOBAL': {0: True}, 'LOCAL': {0: True}, 'SESSION': {0: True}, 'TRANSACTION': {0: True}}
Inherited Members
sqlglot.parser.Parser
Parser
NO_PAREN_FUNCTIONS
STRUCT_TYPE_TOKENS
NESTED_TYPE_TOKENS
ENUM_TYPE_TOKENS
AGGREGATE_TYPE_TOKENS
TYPE_TOKENS
SIGNED_TO_UNSIGNED_TYPE_TOKEN
SUBQUERY_PREDICATES
RESERVED_TOKENS
DB_CREATABLES
CREATABLES
ALTERABLES
INTERVAL_VARS
ALIAS_TOKENS
ARRAY_CONSTRUCTORS
COMMENT_TABLE_ALIAS_TOKENS
UPDATE_ALIAS_TOKENS
TRIM_TYPES
FUNC_TOKENS
CONJUNCTION
ASSIGNMENT
DISJUNCTION
EQUALITY
COMPARISON
TERM
FACTOR
TIMES
TIMESTAMPS
SET_OPERATIONS
JOIN_METHODS
JOIN_SIDES
JOIN_KINDS
JOIN_HINTS
LAMBDAS
COLUMN_OPERATORS
EXPRESSION_PARSERS
UNARY_PARSERS
STRING_PARSERS
NUMERIC_PARSERS
PRIMARY_PARSERS
PROPERTY_PARSERS
CONSTRAINT_PARSERS
ALTER_PARSERS
ALTER_ALTER_PARSERS
SCHEMA_UNNAMED_CONSTRAINTS
INVALID_FUNC_NAME_TOKENS
KEY_VALUE_DEFINITIONS
QUERY_MODIFIER_PARSERS
SET_PARSERS
SHOW_PARSERS
TYPE_LITERAL_PARSERS
DDL_SELECT_TOKENS
PRE_VOLATILE_TOKENS
TRANSACTION_KIND
TRANSACTION_CHARACTERISTICS
CONFLICT_ACTIONS
CREATE_SEQUENCE
ISOLATED_LOADING_OPTIONS
USABLES
CAST_ACTIONS
SCHEMA_BINDING_OPTIONS
PROCEDURE_OPTIONS
EXECUTE_AS_OPTIONS
KEY_CONSTRAINT_OPTIONS
INSERT_ALTERNATIVES
CLONE_KEYWORDS
HISTORICAL_DATA_PREFIX
HISTORICAL_DATA_KIND
OPCLASS_FOLLOW_KEYWORDS
OPTYPE_FOLLOW_TOKENS
TABLE_INDEX_HINT_TOKENS
VIEW_ATTRIBUTES
WINDOW_ALIAS_TOKENS
WINDOW_BEFORE_PAREN_TOKENS
WINDOW_SIDES
JSON_KEY_VALUE_SEPARATOR_TOKENS
FETCH_TOKENS
ADD_CONSTRAINT_TOKENS
DISTINCT_TOKENS
NULL_TOKENS
UNNEST_OFFSET_ALIAS_TOKENS
SELECT_START_TOKENS
COPY_INTO_VARLEN_OPTIONS
IS_JSON_PREDICATE_KIND
ODBC_DATETIME_LITERALS
ON_CONDITION_TOKENS
PRIVILEGE_FOLLOW_TOKENS
DESCRIBE_STYLES
ANALYZE_STYLES
ANALYZE_EXPRESSION_PARSERS
PARTITION_KEYWORDS
AMBIGUOUS_ALIAS_TOKENS
OPERATION_MODIFIERS
STRICT_CAST
PREFIXED_PIVOT_COLUMNS
IDENTIFY_PIVOT_STRINGS
LOG_DEFAULTS_TO_LN
ALTER_TABLE_ADD_REQUIRED_FOR_EACH_COLUMN
TABLESAMPLE_CSV
DEFAULT_SAMPLING_METHOD
SET_REQUIRES_ASSIGNMENT_DELIMITER
TRIM_PATTERN_FIRST
STRING_ALIASES
MODIFIERS_ATTACHED_TO_SET_OP
SET_OP_MODIFIERS
NO_PAREN_IF_COMMANDS
JSON_ARROWS_REQUIRE_JSON_TYPE
COLON_IS_VARIANT_EXTRACT
VALUES_FOLLOWED_BY_PAREN
SUPPORTS_IMPLICIT_UNNEST
INTERVAL_SPANS
SUPPORTS_PARTITION_SELECTION
WRAPPED_TRANSFORM_COLUMN_CONSTRAINT
OPTIONAL_ALIAS_TOKEN_CTE
error_level
error_message_context
max_errors
dialect
reset
parse
parse_into
check_errors
raise_error
expression
validate_expression
errors
sql
class DuckDB.Generator(sqlglot.generator.Generator):
 516    class Generator(generator.Generator):
 517        PARAMETER_TOKEN = "$"
 518        NAMED_PLACEHOLDER_TOKEN = "$"
 519        JOIN_HINTS = False
 520        TABLE_HINTS = False
 521        QUERY_HINTS = False
 522        LIMIT_FETCH = "LIMIT"
 523        STRUCT_DELIMITER = ("(", ")")
 524        RENAME_TABLE_WITH_DB = False
 525        NVL2_SUPPORTED = False
 526        SEMI_ANTI_JOIN_WITH_SIDE = False
 527        TABLESAMPLE_KEYWORDS = "USING SAMPLE"
 528        TABLESAMPLE_SEED_KEYWORD = "REPEATABLE"
 529        LAST_DAY_SUPPORTS_DATE_PART = False
 530        JSON_KEY_VALUE_PAIR_SEP = ","
 531        IGNORE_NULLS_IN_FUNC = True
 532        JSON_PATH_BRACKETED_KEY_SUPPORTED = False
 533        SUPPORTS_CREATE_TABLE_LIKE = False
 534        MULTI_ARG_DISTINCT = False
 535        CAN_IMPLEMENT_ARRAY_ANY = True
 536        SUPPORTS_TO_NUMBER = False
 537        COPY_HAS_INTO_KEYWORD = False
 538        STAR_EXCEPT = "EXCLUDE"
 539        PAD_FILL_PATTERN_IS_REQUIRED = True
 540        ARRAY_CONCAT_IS_VAR_LEN = False
 541        ARRAY_SIZE_DIM_REQUIRED = False
 542
 543        TRANSFORMS = {
 544            **generator.Generator.TRANSFORMS,
 545            exp.ApproxDistinct: approx_count_distinct_sql,
 546            exp.Array: inline_array_unless_query,
 547            exp.ArrayFilter: rename_func("LIST_FILTER"),
 548            exp.ArraySort: _array_sort_sql,
 549            exp.ArraySum: rename_func("LIST_SUM"),
 550            exp.BitwiseXor: rename_func("XOR"),
 551            exp.CommentColumnConstraint: no_comment_column_constraint_sql,
 552            exp.CurrentDate: lambda *_: "CURRENT_DATE",
 553            exp.CurrentTime: lambda *_: "CURRENT_TIME",
 554            exp.CurrentTimestamp: lambda *_: "CURRENT_TIMESTAMP",
 555            exp.DayOfMonth: rename_func("DAYOFMONTH"),
 556            exp.DayOfWeek: rename_func("DAYOFWEEK"),
 557            exp.DayOfWeekIso: rename_func("ISODOW"),
 558            exp.DayOfYear: rename_func("DAYOFYEAR"),
 559            exp.DataType: _datatype_sql,
 560            exp.Date: _date_sql,
 561            exp.DateAdd: _date_delta_sql,
 562            exp.DateFromParts: rename_func("MAKE_DATE"),
 563            exp.DateSub: _date_delta_sql,
 564            exp.DateDiff: _date_diff_sql,
 565            exp.DateStrToDate: datestrtodate_sql,
 566            exp.Datetime: no_datetime_sql,
 567            exp.DatetimeSub: _date_delta_sql,
 568            exp.DatetimeAdd: _date_delta_sql,
 569            exp.DateToDi: lambda self,
 570            e: f"CAST(STRFTIME({self.sql(e, 'this')}, {DuckDB.DATEINT_FORMAT}) AS INT)",
 571            exp.Decode: lambda self, e: encode_decode_sql(self, e, "DECODE", replace=False),
 572            exp.DiToDate: lambda self,
 573            e: f"CAST(STRPTIME(CAST({self.sql(e, 'this')} AS TEXT), {DuckDB.DATEINT_FORMAT}) AS DATE)",
 574            exp.Encode: lambda self, e: encode_decode_sql(self, e, "ENCODE", replace=False),
 575            exp.GenerateDateArray: _generate_datetime_array_sql,
 576            exp.GenerateTimestampArray: _generate_datetime_array_sql,
 577            exp.Explode: rename_func("UNNEST"),
 578            exp.IntDiv: lambda self, e: self.binary(e, "//"),
 579            exp.IsInf: rename_func("ISINF"),
 580            exp.IsNan: rename_func("ISNAN"),
 581            exp.JSONBExists: rename_func("JSON_EXISTS"),
 582            exp.JSONExtract: _arrow_json_extract_sql,
 583            exp.JSONExtractArray: _json_extract_value_array_sql,
 584            exp.JSONExtractScalar: _arrow_json_extract_sql,
 585            exp.JSONFormat: _json_format_sql,
 586            exp.JSONValueArray: _json_extract_value_array_sql,
 587            exp.Lateral: explode_to_unnest_sql,
 588            exp.LogicalOr: rename_func("BOOL_OR"),
 589            exp.LogicalAnd: rename_func("BOOL_AND"),
 590            exp.MakeInterval: lambda self, e: no_make_interval_sql(self, e, sep=" "),
 591            exp.MD5Digest: lambda self, e: self.func("UNHEX", self.func("MD5", e.this)),
 592            exp.MonthsBetween: lambda self, e: self.func(
 593                "DATEDIFF",
 594                "'month'",
 595                exp.cast(e.expression, exp.DataType.Type.TIMESTAMP, copy=True),
 596                exp.cast(e.this, exp.DataType.Type.TIMESTAMP, copy=True),
 597            ),
 598            exp.PercentileCont: rename_func("QUANTILE_CONT"),
 599            exp.PercentileDisc: rename_func("QUANTILE_DISC"),
 600            # DuckDB doesn't allow qualified columns inside of PIVOT expressions.
 601            # See: https://github.com/duckdb/duckdb/blob/671faf92411182f81dce42ac43de8bfb05d9909e/src/planner/binder/tableref/bind_pivot.cpp#L61-L62
 602            exp.Pivot: transforms.preprocess([transforms.unqualify_columns]),
 603            exp.RegexpReplace: lambda self, e: self.func(
 604                "REGEXP_REPLACE",
 605                e.this,
 606                e.expression,
 607                e.args.get("replacement"),
 608                e.args.get("modifiers"),
 609            ),
 610            exp.RegexpLike: rename_func("REGEXP_MATCHES"),
 611            exp.RegexpILike: lambda self, e: self.func(
 612                "REGEXP_MATCHES", e.this, e.expression, exp.Literal.string("i")
 613            ),
 614            exp.RegexpSplit: rename_func("STR_SPLIT_REGEX"),
 615            exp.Return: lambda self, e: self.sql(e, "this"),
 616            exp.ReturnsProperty: lambda self, e: "TABLE" if isinstance(e.this, exp.Schema) else "",
 617            exp.Rand: rename_func("RANDOM"),
 618            exp.SHA: rename_func("SHA1"),
 619            exp.SHA2: sha256_sql,
 620            exp.Split: rename_func("STR_SPLIT"),
 621            exp.SortArray: _sort_array_sql,
 622            exp.StrPosition: strposition_sql,
 623            exp.StrToUnix: lambda self, e: self.func(
 624                "EPOCH", self.func("STRPTIME", e.this, self.format_time(e))
 625            ),
 626            exp.Struct: _struct_sql,
 627            exp.Transform: rename_func("LIST_TRANSFORM"),
 628            exp.TimeAdd: _date_delta_sql,
 629            exp.Time: no_time_sql,
 630            exp.TimeDiff: _timediff_sql,
 631            exp.Timestamp: no_timestamp_sql,
 632            exp.TimestampDiff: lambda self, e: self.func(
 633                "DATE_DIFF", exp.Literal.string(e.unit), e.expression, e.this
 634            ),
 635            exp.TimestampTrunc: timestamptrunc_sql(),
 636            exp.TimeStrToDate: lambda self, e: self.sql(exp.cast(e.this, exp.DataType.Type.DATE)),
 637            exp.TimeStrToTime: timestrtotime_sql,
 638            exp.TimeStrToUnix: lambda self, e: self.func(
 639                "EPOCH", exp.cast(e.this, exp.DataType.Type.TIMESTAMP)
 640            ),
 641            exp.TimeToStr: lambda self, e: self.func("STRFTIME", e.this, self.format_time(e)),
 642            exp.TimeToUnix: rename_func("EPOCH"),
 643            exp.TsOrDiToDi: lambda self,
 644            e: f"CAST(SUBSTR(REPLACE(CAST({self.sql(e, 'this')} AS TEXT), '-', ''), 1, 8) AS INT)",
 645            exp.TsOrDsAdd: _date_delta_sql,
 646            exp.TsOrDsDiff: lambda self, e: self.func(
 647                "DATE_DIFF",
 648                f"'{e.args.get('unit') or 'DAY'}'",
 649                exp.cast(e.expression, exp.DataType.Type.TIMESTAMP),
 650                exp.cast(e.this, exp.DataType.Type.TIMESTAMP),
 651            ),
 652            exp.UnixToStr: lambda self, e: self.func(
 653                "STRFTIME", self.func("TO_TIMESTAMP", e.this), self.format_time(e)
 654            ),
 655            exp.DatetimeTrunc: lambda self, e: self.func(
 656                "DATE_TRUNC", unit_to_str(e), exp.cast(e.this, exp.DataType.Type.DATETIME)
 657            ),
 658            exp.UnixToTime: _unix_to_time_sql,
 659            exp.UnixToTimeStr: lambda self, e: f"CAST(TO_TIMESTAMP({self.sql(e, 'this')}) AS TEXT)",
 660            exp.VariancePop: rename_func("VAR_POP"),
 661            exp.WeekOfYear: rename_func("WEEKOFYEAR"),
 662            exp.Xor: bool_xor_sql,
 663            exp.Levenshtein: unsupported_args("ins_cost", "del_cost", "sub_cost", "max_dist")(
 664                rename_func("LEVENSHTEIN")
 665            ),
 666        }
 667
 668        SUPPORTED_JSON_PATH_PARTS = {
 669            exp.JSONPathKey,
 670            exp.JSONPathRoot,
 671            exp.JSONPathSubscript,
 672            exp.JSONPathWildcard,
 673        }
 674
 675        TYPE_MAPPING = {
 676            **generator.Generator.TYPE_MAPPING,
 677            exp.DataType.Type.BINARY: "BLOB",
 678            exp.DataType.Type.BPCHAR: "TEXT",
 679            exp.DataType.Type.CHAR: "TEXT",
 680            exp.DataType.Type.DATETIME: "TIMESTAMP",
 681            exp.DataType.Type.FLOAT: "REAL",
 682            exp.DataType.Type.NCHAR: "TEXT",
 683            exp.DataType.Type.NVARCHAR: "TEXT",
 684            exp.DataType.Type.UINT: "UINTEGER",
 685            exp.DataType.Type.VARBINARY: "BLOB",
 686            exp.DataType.Type.ROWVERSION: "BLOB",
 687            exp.DataType.Type.VARCHAR: "TEXT",
 688            exp.DataType.Type.TIMESTAMPNTZ: "TIMESTAMP",
 689            exp.DataType.Type.TIMESTAMP_S: "TIMESTAMP_S",
 690            exp.DataType.Type.TIMESTAMP_MS: "TIMESTAMP_MS",
 691            exp.DataType.Type.TIMESTAMP_NS: "TIMESTAMP_NS",
 692        }
 693
 694        # https://github.com/duckdb/duckdb/blob/ff7f24fd8e3128d94371827523dae85ebaf58713/third_party/libpg_query/grammar/keywords/reserved_keywords.list#L1-L77
 695        RESERVED_KEYWORDS = {
 696            "array",
 697            "analyse",
 698            "union",
 699            "all",
 700            "when",
 701            "in_p",
 702            "default",
 703            "create_p",
 704            "window",
 705            "asymmetric",
 706            "to",
 707            "else",
 708            "localtime",
 709            "from",
 710            "end_p",
 711            "select",
 712            "current_date",
 713            "foreign",
 714            "with",
 715            "grant",
 716            "session_user",
 717            "or",
 718            "except",
 719            "references",
 720            "fetch",
 721            "limit",
 722            "group_p",
 723            "leading",
 724            "into",
 725            "collate",
 726            "offset",
 727            "do",
 728            "then",
 729            "localtimestamp",
 730            "check_p",
 731            "lateral_p",
 732            "current_role",
 733            "where",
 734            "asc_p",
 735            "placing",
 736            "desc_p",
 737            "user",
 738            "unique",
 739            "initially",
 740            "column",
 741            "both",
 742            "some",
 743            "as",
 744            "any",
 745            "only",
 746            "deferrable",
 747            "null_p",
 748            "current_time",
 749            "true_p",
 750            "table",
 751            "case",
 752            "trailing",
 753            "variadic",
 754            "for",
 755            "on",
 756            "distinct",
 757            "false_p",
 758            "not",
 759            "constraint",
 760            "current_timestamp",
 761            "returning",
 762            "primary",
 763            "intersect",
 764            "having",
 765            "analyze",
 766            "current_user",
 767            "and",
 768            "cast",
 769            "symmetric",
 770            "using",
 771            "order",
 772            "current_catalog",
 773        }
 774
 775        UNWRAPPED_INTERVAL_VALUES = (exp.Literal, exp.Paren)
 776
 777        # DuckDB doesn't generally support CREATE TABLE .. properties
 778        # https://duckdb.org/docs/sql/statements/create_table.html
 779        PROPERTIES_LOCATION = {
 780            prop: exp.Properties.Location.UNSUPPORTED
 781            for prop in generator.Generator.PROPERTIES_LOCATION
 782        }
 783
 784        # There are a few exceptions (e.g. temporary tables) which are supported or
 785        # can be transpiled to DuckDB, so we explicitly override them accordingly
 786        PROPERTIES_LOCATION[exp.LikeProperty] = exp.Properties.Location.POST_SCHEMA
 787        PROPERTIES_LOCATION[exp.TemporaryProperty] = exp.Properties.Location.POST_CREATE
 788        PROPERTIES_LOCATION[exp.ReturnsProperty] = exp.Properties.Location.POST_ALIAS
 789
 790        def fromiso8601timestamp_sql(self, expression: exp.FromISO8601Timestamp) -> str:
 791            return self.sql(exp.cast(expression.this, exp.DataType.Type.TIMESTAMPTZ))
 792
 793        def strtotime_sql(self, expression: exp.StrToTime) -> str:
 794            if expression.args.get("safe"):
 795                formatted_time = self.format_time(expression)
 796                return f"CAST({self.func('TRY_STRPTIME', expression.this, formatted_time)} AS TIMESTAMP)"
 797            return str_to_time_sql(self, expression)
 798
 799        def strtodate_sql(self, expression: exp.StrToDate) -> str:
 800            if expression.args.get("safe"):
 801                formatted_time = self.format_time(expression)
 802                return f"CAST({self.func('TRY_STRPTIME', expression.this, formatted_time)} AS DATE)"
 803            return f"CAST({str_to_time_sql(self, expression)} AS DATE)"
 804
 805        def parsejson_sql(self, expression: exp.ParseJSON) -> str:
 806            arg = expression.this
 807            if expression.args.get("safe"):
 808                return self.sql(exp.case().when(exp.func("json_valid", arg), arg).else_(exp.null()))
 809            return self.func("JSON", arg)
 810
 811        def timefromparts_sql(self, expression: exp.TimeFromParts) -> str:
 812            nano = expression.args.get("nano")
 813            if nano is not None:
 814                expression.set(
 815                    "sec", expression.args["sec"] + nano.pop() / exp.Literal.number(1000000000.0)
 816                )
 817
 818            return rename_func("MAKE_TIME")(self, expression)
 819
 820        def timestampfromparts_sql(self, expression: exp.TimestampFromParts) -> str:
 821            sec = expression.args["sec"]
 822
 823            milli = expression.args.get("milli")
 824            if milli is not None:
 825                sec += milli.pop() / exp.Literal.number(1000.0)
 826
 827            nano = expression.args.get("nano")
 828            if nano is not None:
 829                sec += nano.pop() / exp.Literal.number(1000000000.0)
 830
 831            if milli or nano:
 832                expression.set("sec", sec)
 833
 834            return rename_func("MAKE_TIMESTAMP")(self, expression)
 835
 836        def tablesample_sql(
 837            self,
 838            expression: exp.TableSample,
 839            tablesample_keyword: t.Optional[str] = None,
 840        ) -> str:
 841            if not isinstance(expression.parent, exp.Select):
 842                # This sample clause only applies to a single source, not the entire resulting relation
 843                tablesample_keyword = "TABLESAMPLE"
 844
 845            if expression.args.get("size"):
 846                method = expression.args.get("method")
 847                if method and method.name.upper() != "RESERVOIR":
 848                    self.unsupported(
 849                        f"Sampling method {method} is not supported with a discrete sample count, "
 850                        "defaulting to reservoir sampling"
 851                    )
 852                    expression.set("method", exp.var("RESERVOIR"))
 853
 854            return super().tablesample_sql(expression, tablesample_keyword=tablesample_keyword)
 855
 856        def interval_sql(self, expression: exp.Interval) -> str:
 857            multiplier: t.Optional[int] = None
 858            unit = expression.text("unit").lower()
 859
 860            if unit.startswith("week"):
 861                multiplier = 7
 862            if unit.startswith("quarter"):
 863                multiplier = 90
 864
 865            if multiplier:
 866                return f"({multiplier} * {super().interval_sql(exp.Interval(this=expression.this, unit=exp.var('DAY')))})"
 867
 868            return super().interval_sql(expression)
 869
 870        def columndef_sql(self, expression: exp.ColumnDef, sep: str = " ") -> str:
 871            if isinstance(expression.parent, exp.UserDefinedFunction):
 872                return self.sql(expression, "this")
 873            return super().columndef_sql(expression, sep)
 874
 875        def join_sql(self, expression: exp.Join) -> str:
 876            if (
 877                expression.side == "LEFT"
 878                and not expression.args.get("on")
 879                and isinstance(expression.this, exp.Unnest)
 880            ):
 881                # Some dialects support `LEFT JOIN UNNEST(...)` without an explicit ON clause
 882                # DuckDB doesn't, but we can just add a dummy ON clause that is always true
 883                return super().join_sql(expression.on(exp.true()))
 884
 885            return super().join_sql(expression)
 886
 887        def generateseries_sql(self, expression: exp.GenerateSeries) -> str:
 888            # GENERATE_SERIES(a, b) -> [a, b], RANGE(a, b) -> [a, b)
 889            if expression.args.get("is_end_exclusive"):
 890                return rename_func("RANGE")(self, expression)
 891
 892            return self.function_fallback_sql(expression)
 893
 894        def bracket_sql(self, expression: exp.Bracket) -> str:
 895            this = expression.this
 896            if isinstance(this, exp.Array):
 897                this.replace(exp.paren(this))
 898
 899            bracket = super().bracket_sql(expression)
 900
 901            if not expression.args.get("returns_list_for_maps"):
 902                if not this.type:
 903                    from sqlglot.optimizer.annotate_types import annotate_types
 904
 905                    this = annotate_types(this)
 906
 907                if this.is_type(exp.DataType.Type.MAP):
 908                    bracket = f"({bracket})[1]"
 909
 910            return bracket
 911
 912        def withingroup_sql(self, expression: exp.WithinGroup) -> str:
 913            expression_sql = self.sql(expression, "expression")
 914
 915            func = expression.this
 916            if isinstance(func, exp.PERCENTILES):
 917                # Make the order key the first arg and slide the fraction to the right
 918                # https://duckdb.org/docs/sql/aggregates#ordered-set-aggregate-functions
 919                order_col = expression.find(exp.Ordered)
 920                if order_col:
 921                    func.set("expression", func.this)
 922                    func.set("this", order_col.this)
 923
 924            this = self.sql(expression, "this").rstrip(")")
 925
 926            return f"{this}{expression_sql})"
 927
 928        def length_sql(self, expression: exp.Length) -> str:
 929            arg = expression.this
 930
 931            # Dialects like BQ and Snowflake also accept binary values as args, so
 932            # DDB will attempt to infer the type or resort to case/when resolution
 933            if not expression.args.get("binary") or arg.is_string:
 934                return self.func("LENGTH", arg)
 935
 936            if not arg.type:
 937                from sqlglot.optimizer.annotate_types import annotate_types
 938
 939                arg = annotate_types(arg)
 940
 941            if arg.is_type(*exp.DataType.TEXT_TYPES):
 942                return self.func("LENGTH", arg)
 943
 944            # We need these casts to make duckdb's static type checker happy
 945            blob = exp.cast(arg, exp.DataType.Type.VARBINARY)
 946            varchar = exp.cast(arg, exp.DataType.Type.VARCHAR)
 947
 948            case = (
 949                exp.case(self.func("TYPEOF", arg))
 950                .when(
 951                    "'VARCHAR'", exp.Anonymous(this="LENGTH", expressions=[varchar])
 952                )  # anonymous to break length_sql recursion
 953                .when("'BLOB'", self.func("OCTET_LENGTH", blob))
 954            )
 955
 956            return self.sql(case)
 957
 958        def objectinsert_sql(self, expression: exp.ObjectInsert) -> str:
 959            this = expression.this
 960            key = expression.args.get("key")
 961            key_sql = key.name if isinstance(key, exp.Expression) else ""
 962            value_sql = self.sql(expression, "value")
 963
 964            kv_sql = f"{key_sql} := {value_sql}"
 965
 966            # If the input struct is empty e.g. transpiling OBJECT_INSERT(OBJECT_CONSTRUCT(), key, value) from Snowflake
 967            # then we can generate STRUCT_PACK which will build it since STRUCT_INSERT({}, key := value) is not valid DuckDB
 968            if isinstance(this, exp.Struct) and not this.expressions:
 969                return self.func("STRUCT_PACK", kv_sql)
 970
 971            return self.func("STRUCT_INSERT", this, kv_sql)
 972
 973        def unnest_sql(self, expression: exp.Unnest) -> str:
 974            explode_array = expression.args.get("explode_array")
 975            if explode_array:
 976                # In BigQuery, UNNESTing a nested array leads to explosion of the top-level array & struct
 977                # This is transpiled to DDB by transforming "FROM UNNEST(...)" to "FROM (SELECT UNNEST(..., max_depth => 2))"
 978                expression.expressions.append(
 979                    exp.Kwarg(this=exp.var("max_depth"), expression=exp.Literal.number(2))
 980                )
 981
 982                # If BQ's UNNEST is aliased, we transform it from a column alias to a table alias in DDB
 983                alias = expression.args.get("alias")
 984                if alias:
 985                    expression.set("alias", None)
 986                    alias = exp.TableAlias(this=seq_get(alias.args.get("columns"), 0))
 987
 988                unnest_sql = super().unnest_sql(expression)
 989                select = exp.Select(expressions=[unnest_sql]).subquery(alias)
 990                return self.sql(select)
 991
 992            return super().unnest_sql(expression)
 993
 994        def ignorenulls_sql(self, expression: exp.IgnoreNulls) -> str:
 995            if isinstance(expression.this, WINDOW_FUNCS_WITH_IGNORE_NULLS):
 996                # DuckDB should render IGNORE NULLS only for the general-purpose
 997                # window functions that accept it e.g. FIRST_VALUE(... IGNORE NULLS) OVER (...)
 998                return super().ignorenulls_sql(expression)
 999
1000            return self.sql(expression, "this")
1001
1002        def arraytostring_sql(self, expression: exp.ArrayToString) -> str:
1003            this = self.sql(expression, "this")
1004            null_text = self.sql(expression, "null")
1005
1006            if null_text:
1007                this = f"LIST_TRANSFORM({this}, x -> COALESCE(x, {null_text}))"
1008
1009            return self.func("ARRAY_TO_STRING", this, expression.expression)
1010
1011        @unsupported_args("position", "occurrence")
1012        def regexpextract_sql(self, expression: exp.RegexpExtract) -> str:
1013            group = expression.args.get("group")
1014            params = expression.args.get("parameters")
1015
1016            # Do not render group if there is no following argument,
1017            # and it's the default value for this dialect
1018            if (
1019                not params
1020                and group
1021                and group.name == str(self.dialect.REGEXP_EXTRACT_DEFAULT_GROUP)
1022            ):
1023                group = None
1024            return self.func(
1025                "REGEXP_EXTRACT", expression.this, expression.expression, group, params
1026            )
1027
1028        @unsupported_args("culture")
1029        def numbertostr_sql(self, expression: exp.NumberToStr) -> str:
1030            fmt = expression.args.get("format")
1031            if fmt and fmt.is_int:
1032                return self.func("FORMAT", f"'{{:,.{fmt.name}f}}'", expression.this)
1033
1034            self.unsupported("Only integer formats are supported by NumberToStr")
1035            return self.function_fallback_sql(expression)

Generator converts a given syntax tree to the corresponding SQL string.

Arguments:
  • pretty: Whether to format the produced SQL string. Default: False.
  • identify: Determines when an identifier should be quoted. Possible values are: False (default): Never quote, except in cases where it's mandatory by the dialect. True or 'always': Always quote. 'safe': Only quote identifiers that are case insensitive.
  • normalize: Whether to normalize identifiers to lowercase. Default: False.
  • pad: The pad size in a formatted string. For example, this affects the indentation of a projection in a query, relative to its nesting level. Default: 2.
  • indent: The indentation size in a formatted string. For example, this affects the indentation of subqueries and filters under a WHERE clause. Default: 2.
  • normalize_functions: How to normalize function names. Possible values are: "upper" or True (default): Convert names to uppercase. "lower": Convert names to lowercase. False: Disables function name normalization.
  • unsupported_level: Determines the generator's behavior when it encounters unsupported expressions. Default ErrorLevel.WARN.
  • max_unsupported: Maximum number of unsupported messages to include in a raised UnsupportedError. This is only relevant if unsupported_level is ErrorLevel.RAISE. Default: 3
  • leading_comma: Whether the comma is leading or trailing in select expressions. This is only relevant when generating in pretty mode. Default: False
  • max_text_width: The max number of characters in a segment before creating new lines in pretty mode. The default is on the smaller end because the length only represents a segment and not the true line length. Default: 80
  • comments: Whether to preserve comments in the output SQL code. Default: True
PARAMETER_TOKEN = '$'
NAMED_PLACEHOLDER_TOKEN = '$'
JOIN_HINTS = False
TABLE_HINTS = False
QUERY_HINTS = False
LIMIT_FETCH = 'LIMIT'
STRUCT_DELIMITER = ('(', ')')
RENAME_TABLE_WITH_DB = False
NVL2_SUPPORTED = False
SEMI_ANTI_JOIN_WITH_SIDE = False
TABLESAMPLE_KEYWORDS = 'USING SAMPLE'
TABLESAMPLE_SEED_KEYWORD = 'REPEATABLE'
LAST_DAY_SUPPORTS_DATE_PART = False
JSON_KEY_VALUE_PAIR_SEP = ','
IGNORE_NULLS_IN_FUNC = True
JSON_PATH_BRACKETED_KEY_SUPPORTED = False
SUPPORTS_CREATE_TABLE_LIKE = False
MULTI_ARG_DISTINCT = False
CAN_IMPLEMENT_ARRAY_ANY = True
SUPPORTS_TO_NUMBER = False
COPY_HAS_INTO_KEYWORD = False
STAR_EXCEPT = 'EXCLUDE'
PAD_FILL_PATTERN_IS_REQUIRED = True
ARRAY_CONCAT_IS_VAR_LEN = False
ARRAY_SIZE_DIM_REQUIRED = False
TRANSFORMS = {<class 'sqlglot.expressions.JSONPathKey'>: <function <lambda>>, <class 'sqlglot.expressions.JSONPathRoot'>: <function <lambda>>, <class 'sqlglot.expressions.JSONPathSubscript'>: <function <lambda>>, <class 'sqlglot.expressions.JSONPathWildcard'>: <function <lambda>>, <class 'sqlglot.expressions.AllowedValuesProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.AnalyzeColumns'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.AnalyzeWith'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.ArrayContainsAll'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.ArrayOverlaps'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.AutoRefreshProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.BackupProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.CaseSpecificColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.Ceil'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.CharacterSetColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.CharacterSetProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.ClusteredColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.CollateColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.CommentColumnConstraint'>: <function no_comment_column_constraint_sql>, <class 'sqlglot.expressions.ConnectByRoot'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.CopyGrantsProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.DateFormatColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.DefaultColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.DynamicProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.EmptyProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.EncodeColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.EphemeralColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.ExcludeColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.ExecuteAsProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.Except'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.ExternalProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.Floor'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.GlobalProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.HeapProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.IcebergProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.InheritsProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.InlineLengthColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.InputModelProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.Intersect'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.IntervalSpan'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.Int64'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.LanguageProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.LocationProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.LogProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.MaterializedProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.NonClusteredColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.NoPrimaryIndexProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.NotForReplicationColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.OnCommitProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.OnProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.OnUpdateColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.Operator'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.OutputModelProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.PathColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.PivotAny'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.ProjectionPolicyColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.RemoteWithConnectionModelProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.ReturnsProperty'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.SampleProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.SecureProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.SecurityProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.SetConfigProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.SetProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.SettingsProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.SharingProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.SqlReadWriteProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.SqlSecurityProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.StabilityProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.Stream'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.StreamingTableProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.StrictProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.SwapTable'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.TemporaryProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.Tags'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.TitleColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.ToMap'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.ToTableProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.TransformModelProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.TransientProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.Union'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.UnloggedProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.UsingData'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.Uuid'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.UppercaseColumnConstraint'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.VarMap'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.ViewAttributeProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.VolatileProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.WithJournalTableProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.WithProcedureOptions'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.WithSchemaBindingProperty'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.WithOperator'>: <function Generator.<lambda>>, <class 'sqlglot.expressions.ApproxDistinct'>: <function approx_count_distinct_sql>, <class 'sqlglot.expressions.Array'>: <function inline_array_unless_query>, <class 'sqlglot.expressions.ArrayFilter'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.ArraySort'>: <function _array_sort_sql>, <class 'sqlglot.expressions.ArraySum'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.BitwiseXor'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.CurrentDate'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.CurrentTime'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.CurrentTimestamp'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.DayOfMonth'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.DayOfWeek'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.DayOfWeekIso'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.DayOfYear'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.DataType'>: <function _datatype_sql>, <class 'sqlglot.expressions.Date'>: <function _date_sql>, <class 'sqlglot.expressions.DateAdd'>: <function _date_delta_sql>, <class 'sqlglot.expressions.DateFromParts'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.DateSub'>: <function _date_delta_sql>, <class 'sqlglot.expressions.DateDiff'>: <function _date_diff_sql>, <class 'sqlglot.expressions.DateStrToDate'>: <function datestrtodate_sql>, <class 'sqlglot.expressions.Datetime'>: <function no_datetime_sql>, <class 'sqlglot.expressions.DatetimeSub'>: <function _date_delta_sql>, <class 'sqlglot.expressions.DatetimeAdd'>: <function _date_delta_sql>, <class 'sqlglot.expressions.DateToDi'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.Decode'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.DiToDate'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.Encode'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.GenerateDateArray'>: <function _generate_datetime_array_sql>, <class 'sqlglot.expressions.GenerateTimestampArray'>: <function _generate_datetime_array_sql>, <class 'sqlglot.expressions.Explode'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.IntDiv'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.IsInf'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.IsNan'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.JSONBExists'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.JSONExtract'>: <function _arrow_json_extract_sql>, <class 'sqlglot.expressions.JSONExtractArray'>: <function _json_extract_value_array_sql>, <class 'sqlglot.expressions.JSONExtractScalar'>: <function _arrow_json_extract_sql>, <class 'sqlglot.expressions.JSONFormat'>: <function _json_format_sql>, <class 'sqlglot.expressions.JSONValueArray'>: <function _json_extract_value_array_sql>, <class 'sqlglot.expressions.Lateral'>: <function explode_to_unnest_sql>, <class 'sqlglot.expressions.LogicalOr'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.LogicalAnd'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.MakeInterval'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.MD5Digest'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.MonthsBetween'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.PercentileCont'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.PercentileDisc'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.Pivot'>: <function preprocess.<locals>._to_sql>, <class 'sqlglot.expressions.RegexpReplace'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.RegexpLike'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.RegexpILike'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.RegexpSplit'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.Return'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.Rand'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.SHA'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.SHA2'>: <function sha256_sql>, <class 'sqlglot.expressions.Split'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.SortArray'>: <function _sort_array_sql>, <class 'sqlglot.expressions.StrPosition'>: <function strposition_sql>, <class 'sqlglot.expressions.StrToUnix'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.Struct'>: <function _struct_sql>, <class 'sqlglot.expressions.Transform'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.TimeAdd'>: <function _date_delta_sql>, <class 'sqlglot.expressions.Time'>: <function no_time_sql>, <class 'sqlglot.expressions.TimeDiff'>: <function _timediff_sql>, <class 'sqlglot.expressions.Timestamp'>: <function no_timestamp_sql>, <class 'sqlglot.expressions.TimestampDiff'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.TimestampTrunc'>: <function timestamptrunc_sql.<locals>._timestamptrunc_sql>, <class 'sqlglot.expressions.TimeStrToDate'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.TimeStrToTime'>: <function timestrtotime_sql>, <class 'sqlglot.expressions.TimeStrToUnix'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.TimeToStr'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.TimeToUnix'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.TsOrDiToDi'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.TsOrDsAdd'>: <function _date_delta_sql>, <class 'sqlglot.expressions.TsOrDsDiff'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.UnixToStr'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.DatetimeTrunc'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.UnixToTime'>: <function _unix_to_time_sql>, <class 'sqlglot.expressions.UnixToTimeStr'>: <function DuckDB.Generator.<lambda>>, <class 'sqlglot.expressions.VariancePop'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.WeekOfYear'>: <function rename_func.<locals>.<lambda>>, <class 'sqlglot.expressions.Xor'>: <function bool_xor_sql>, <class 'sqlglot.expressions.Levenshtein'>: <function rename_func.<locals>.<lambda>>}
TYPE_MAPPING = {<Type.DATETIME2: 'DATETIME2'>: 'TIMESTAMP', <Type.NCHAR: 'NCHAR'>: 'TEXT', <Type.NVARCHAR: 'NVARCHAR'>: 'TEXT', <Type.MEDIUMTEXT: 'MEDIUMTEXT'>: 'TEXT', <Type.LONGTEXT: 'LONGTEXT'>: 'TEXT', <Type.TINYTEXT: 'TINYTEXT'>: 'TEXT', <Type.MEDIUMBLOB: 'MEDIUMBLOB'>: 'BLOB', <Type.LONGBLOB: 'LONGBLOB'>: 'BLOB', <Type.TINYBLOB: 'TINYBLOB'>: 'BLOB', <Type.INET: 'INET'>: 'INET', <Type.ROWVERSION: 'ROWVERSION'>: 'BLOB', <Type.SMALLDATETIME: 'SMALLDATETIME'>: 'TIMESTAMP', <Type.BINARY: 'BINARY'>: 'BLOB', <Type.BPCHAR: 'BPCHAR'>: 'TEXT', <Type.CHAR: 'CHAR'>: 'TEXT', <Type.DATETIME: 'DATETIME'>: 'TIMESTAMP', <Type.FLOAT: 'FLOAT'>: 'REAL', <Type.UINT: 'UINT'>: 'UINTEGER', <Type.VARBINARY: 'VARBINARY'>: 'BLOB', <Type.VARCHAR: 'VARCHAR'>: 'TEXT', <Type.TIMESTAMPNTZ: 'TIMESTAMPNTZ'>: 'TIMESTAMP', <Type.TIMESTAMP_S: 'TIMESTAMP_S'>: 'TIMESTAMP_S', <Type.TIMESTAMP_MS: 'TIMESTAMP_MS'>: 'TIMESTAMP_MS', <Type.TIMESTAMP_NS: 'TIMESTAMP_NS'>: 'TIMESTAMP_NS'}
RESERVED_KEYWORDS = {'with', 'asc_p', 'intersect', 'both', 'as', 'current_date', 'else', 'session_user', 'analyse', 'into', 'and', 'all', 'select', 'references', 'user', 'desc_p', 'some', 'null_p', 'using', 'analyze', 'localtimestamp', 'foreign', 'any', 'collate', 'from', 'then', 'having', 'array', 'table', 'fetch', 'column', 'constraint', 'limit', 'placing', 'default', 'current_user', 'returning', 'in_p', 'deferrable', 'true_p', 'current_role', 'when', 'leading', 'current_catalog', 'offset', 'union', 'only', 'lateral_p', 'end_p', 'distinct', 'where', 'current_time', 'on', 'variadic', 'asymmetric', 'or', 'for', 'false_p', 'symmetric', 'grant', 'window', 'order', 'initially', 'current_timestamp', 'to', 'except', 'case', 'create_p', 'unique', 'trailing', 'localtime', 'not', 'do', 'primary', 'check_p', 'group_p', 'cast'}
UNWRAPPED_INTERVAL_VALUES = (<class 'sqlglot.expressions.Literal'>, <class 'sqlglot.expressions.Paren'>)
PROPERTIES_LOCATION = {<class 'sqlglot.expressions.AllowedValuesProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.AlgorithmProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.AutoIncrementProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.AutoRefreshProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.BackupProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.BlockCompressionProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.CharacterSetProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.ChecksumProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.CollateProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.CopyGrantsProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.Cluster'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.ClusteredByProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.DistributedByProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.DuplicateKeyProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.DataBlocksizeProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.DataDeletionProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.DefinerProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.DictRange'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.DictProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.DynamicProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.DistKeyProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.DistStyleProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.EmptyProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.EncodeProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.EngineProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.ExecuteAsProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.ExternalProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.FallbackProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.FileFormatProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.FreespaceProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.GlobalProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.HeapProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.InheritsProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.IcebergProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.IncludeProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.InputModelProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.IsolatedLoadingProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.JournalProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.LanguageProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.LikeProperty'>: <Location.POST_SCHEMA: 'POST_SCHEMA'>, <class 'sqlglot.expressions.LocationProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.LockProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.LockingProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.LogProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.MaterializedProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.MergeBlockRatioProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.NoPrimaryIndexProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.OnProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.OnCommitProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.Order'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.OutputModelProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.PartitionedByProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.PartitionedOfProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.PrimaryKey'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.Property'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.RemoteWithConnectionModelProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.ReturnsProperty'>: <Location.POST_ALIAS: 'POST_ALIAS'>, <class 'sqlglot.expressions.RowFormatProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.RowFormatDelimitedProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.RowFormatSerdeProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.SampleProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.SchemaCommentProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.SecureProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.SecurityProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.SerdeProperties'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.Set'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.SettingsProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.SetProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.SetConfigProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.SharingProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.SequenceProperties'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.SortKeyProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.SqlReadWriteProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.SqlSecurityProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.StabilityProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.StreamingTableProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.StrictProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.Tags'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.TemporaryProperty'>: <Location.POST_CREATE: 'POST_CREATE'>, <class 'sqlglot.expressions.ToTableProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.TransientProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.TransformModelProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.MergeTreeTTL'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.UnloggedProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.ViewAttributeProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.VolatileProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.WithDataProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.WithJournalTableProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.WithProcedureOptions'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.WithSchemaBindingProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>, <class 'sqlglot.expressions.WithSystemVersioningProperty'>: <Location.UNSUPPORTED: 'UNSUPPORTED'>}
def fromiso8601timestamp_sql(self, expression: sqlglot.expressions.FromISO8601Timestamp) -> str:
790        def fromiso8601timestamp_sql(self, expression: exp.FromISO8601Timestamp) -> str:
791            return self.sql(exp.cast(expression.this, exp.DataType.Type.TIMESTAMPTZ))
def strtotime_sql(self, expression: sqlglot.expressions.StrToTime) -> str:
793        def strtotime_sql(self, expression: exp.StrToTime) -> str:
794            if expression.args.get("safe"):
795                formatted_time = self.format_time(expression)
796                return f"CAST({self.func('TRY_STRPTIME', expression.this, formatted_time)} AS TIMESTAMP)"
797            return str_to_time_sql(self, expression)
def strtodate_sql(self, expression: sqlglot.expressions.StrToDate) -> str:
799        def strtodate_sql(self, expression: exp.StrToDate) -> str:
800            if expression.args.get("safe"):
801                formatted_time = self.format_time(expression)
802                return f"CAST({self.func('TRY_STRPTIME', expression.this, formatted_time)} AS DATE)"
803            return f"CAST({str_to_time_sql(self, expression)} AS DATE)"
def parsejson_sql(self, expression: sqlglot.expressions.ParseJSON) -> str:
805        def parsejson_sql(self, expression: exp.ParseJSON) -> str:
806            arg = expression.this
807            if expression.args.get("safe"):
808                return self.sql(exp.case().when(exp.func("json_valid", arg), arg).else_(exp.null()))
809            return self.func("JSON", arg)
def timefromparts_sql(self, expression: sqlglot.expressions.TimeFromParts) -> str:
811        def timefromparts_sql(self, expression: exp.TimeFromParts) -> str:
812            nano = expression.args.get("nano")
813            if nano is not None:
814                expression.set(
815                    "sec", expression.args["sec"] + nano.pop() / exp.Literal.number(1000000000.0)
816                )
817
818            return rename_func("MAKE_TIME")(self, expression)
def timestampfromparts_sql(self, expression: sqlglot.expressions.TimestampFromParts) -> str:
820        def timestampfromparts_sql(self, expression: exp.TimestampFromParts) -> str:
821            sec = expression.args["sec"]
822
823            milli = expression.args.get("milli")
824            if milli is not None:
825                sec += milli.pop() / exp.Literal.number(1000.0)
826
827            nano = expression.args.get("nano")
828            if nano is not None:
829                sec += nano.pop() / exp.Literal.number(1000000000.0)
830
831            if milli or nano:
832                expression.set("sec", sec)
833
834            return rename_func("MAKE_TIMESTAMP")(self, expression)
def tablesample_sql( self, expression: sqlglot.expressions.TableSample, tablesample_keyword: Optional[str] = None) -> str:
836        def tablesample_sql(
837            self,
838            expression: exp.TableSample,
839            tablesample_keyword: t.Optional[str] = None,
840        ) -> str:
841            if not isinstance(expression.parent, exp.Select):
842                # This sample clause only applies to a single source, not the entire resulting relation
843                tablesample_keyword = "TABLESAMPLE"
844
845            if expression.args.get("size"):
846                method = expression.args.get("method")
847                if method and method.name.upper() != "RESERVOIR":
848                    self.unsupported(
849                        f"Sampling method {method} is not supported with a discrete sample count, "
850                        "defaulting to reservoir sampling"
851                    )
852                    expression.set("method", exp.var("RESERVOIR"))
853
854            return super().tablesample_sql(expression, tablesample_keyword=tablesample_keyword)
def interval_sql(self, expression: sqlglot.expressions.Interval) -> str:
856        def interval_sql(self, expression: exp.Interval) -> str:
857            multiplier: t.Optional[int] = None
858            unit = expression.text("unit").lower()
859
860            if unit.startswith("week"):
861                multiplier = 7
862            if unit.startswith("quarter"):
863                multiplier = 90
864
865            if multiplier:
866                return f"({multiplier} * {super().interval_sql(exp.Interval(this=expression.this, unit=exp.var('DAY')))})"
867
868            return super().interval_sql(expression)
def columndef_sql(self, expression: sqlglot.expressions.ColumnDef, sep: str = ' ') -> str:
870        def columndef_sql(self, expression: exp.ColumnDef, sep: str = " ") -> str:
871            if isinstance(expression.parent, exp.UserDefinedFunction):
872                return self.sql(expression, "this")
873            return super().columndef_sql(expression, sep)
def join_sql(self, expression: sqlglot.expressions.Join) -> str:
875        def join_sql(self, expression: exp.Join) -> str:
876            if (
877                expression.side == "LEFT"
878                and not expression.args.get("on")
879                and isinstance(expression.this, exp.Unnest)
880            ):
881                # Some dialects support `LEFT JOIN UNNEST(...)` without an explicit ON clause
882                # DuckDB doesn't, but we can just add a dummy ON clause that is always true
883                return super().join_sql(expression.on(exp.true()))
884
885            return super().join_sql(expression)
def generateseries_sql(self, expression: sqlglot.expressions.GenerateSeries) -> str:
887        def generateseries_sql(self, expression: exp.GenerateSeries) -> str:
888            # GENERATE_SERIES(a, b) -> [a, b], RANGE(a, b) -> [a, b)
889            if expression.args.get("is_end_exclusive"):
890                return rename_func("RANGE")(self, expression)
891
892            return self.function_fallback_sql(expression)
def bracket_sql(self, expression: sqlglot.expressions.Bracket) -> str:
894        def bracket_sql(self, expression: exp.Bracket) -> str:
895            this = expression.this
896            if isinstance(this, exp.Array):
897                this.replace(exp.paren(this))
898
899            bracket = super().bracket_sql(expression)
900
901            if not expression.args.get("returns_list_for_maps"):
902                if not this.type:
903                    from sqlglot.optimizer.annotate_types import annotate_types
904
905                    this = annotate_types(this)
906
907                if this.is_type(exp.DataType.Type.MAP):
908                    bracket = f"({bracket})[1]"
909
910            return bracket
def withingroup_sql(self, expression: sqlglot.expressions.WithinGroup) -> str:
912        def withingroup_sql(self, expression: exp.WithinGroup) -> str:
913            expression_sql = self.sql(expression, "expression")
914
915            func = expression.this
916            if isinstance(func, exp.PERCENTILES):
917                # Make the order key the first arg and slide the fraction to the right
918                # https://duckdb.org/docs/sql/aggregates#ordered-set-aggregate-functions
919                order_col = expression.find(exp.Ordered)
920                if order_col:
921                    func.set("expression", func.this)
922                    func.set("this", order_col.this)
923
924            this = self.sql(expression, "this").rstrip(")")
925
926            return f"{this}{expression_sql})"
def length_sql(self, expression: sqlglot.expressions.Length) -> str:
928        def length_sql(self, expression: exp.Length) -> str:
929            arg = expression.this
930
931            # Dialects like BQ and Snowflake also accept binary values as args, so
932            # DDB will attempt to infer the type or resort to case/when resolution
933            if not expression.args.get("binary") or arg.is_string:
934                return self.func("LENGTH", arg)
935
936            if not arg.type:
937                from sqlglot.optimizer.annotate_types import annotate_types
938
939                arg = annotate_types(arg)
940
941            if arg.is_type(*exp.DataType.TEXT_TYPES):
942                return self.func("LENGTH", arg)
943
944            # We need these casts to make duckdb's static type checker happy
945            blob = exp.cast(arg, exp.DataType.Type.VARBINARY)
946            varchar = exp.cast(arg, exp.DataType.Type.VARCHAR)
947
948            case = (
949                exp.case(self.func("TYPEOF", arg))
950                .when(
951                    "'VARCHAR'", exp.Anonymous(this="LENGTH", expressions=[varchar])
952                )  # anonymous to break length_sql recursion
953                .when("'BLOB'", self.func("OCTET_LENGTH", blob))
954            )
955
956            return self.sql(case)
def objectinsert_sql(self, expression: sqlglot.expressions.ObjectInsert) -> str:
958        def objectinsert_sql(self, expression: exp.ObjectInsert) -> str:
959            this = expression.this
960            key = expression.args.get("key")
961            key_sql = key.name if isinstance(key, exp.Expression) else ""
962            value_sql = self.sql(expression, "value")
963
964            kv_sql = f"{key_sql} := {value_sql}"
965
966            # If the input struct is empty e.g. transpiling OBJECT_INSERT(OBJECT_CONSTRUCT(), key, value) from Snowflake
967            # then we can generate STRUCT_PACK which will build it since STRUCT_INSERT({}, key := value) is not valid DuckDB
968            if isinstance(this, exp.Struct) and not this.expressions:
969                return self.func("STRUCT_PACK", kv_sql)
970
971            return self.func("STRUCT_INSERT", this, kv_sql)
def unnest_sql(self, expression: sqlglot.expressions.Unnest) -> str:
973        def unnest_sql(self, expression: exp.Unnest) -> str:
974            explode_array = expression.args.get("explode_array")
975            if explode_array:
976                # In BigQuery, UNNESTing a nested array leads to explosion of the top-level array & struct
977                # This is transpiled to DDB by transforming "FROM UNNEST(...)" to "FROM (SELECT UNNEST(..., max_depth => 2))"
978                expression.expressions.append(
979                    exp.Kwarg(this=exp.var("max_depth"), expression=exp.Literal.number(2))
980                )
981
982                # If BQ's UNNEST is aliased, we transform it from a column alias to a table alias in DDB
983                alias = expression.args.get("alias")
984                if alias:
985                    expression.set("alias", None)
986                    alias = exp.TableAlias(this=seq_get(alias.args.get("columns"), 0))
987
988                unnest_sql = super().unnest_sql(expression)
989                select = exp.Select(expressions=[unnest_sql]).subquery(alias)
990                return self.sql(select)
991
992            return super().unnest_sql(expression)
def ignorenulls_sql(self, expression: sqlglot.expressions.IgnoreNulls) -> str:
 994        def ignorenulls_sql(self, expression: exp.IgnoreNulls) -> str:
 995            if isinstance(expression.this, WINDOW_FUNCS_WITH_IGNORE_NULLS):
 996                # DuckDB should render IGNORE NULLS only for the general-purpose
 997                # window functions that accept it e.g. FIRST_VALUE(... IGNORE NULLS) OVER (...)
 998                return super().ignorenulls_sql(expression)
 999
1000            return self.sql(expression, "this")
def arraytostring_sql(self, expression: sqlglot.expressions.ArrayToString) -> str:
1002        def arraytostring_sql(self, expression: exp.ArrayToString) -> str:
1003            this = self.sql(expression, "this")
1004            null_text = self.sql(expression, "null")
1005
1006            if null_text:
1007                this = f"LIST_TRANSFORM({this}, x -> COALESCE(x, {null_text}))"
1008
1009            return self.func("ARRAY_TO_STRING", this, expression.expression)
@unsupported_args('position', 'occurrence')
def regexpextract_sql(self, expression: sqlglot.expressions.RegexpExtract) -> str:
1011        @unsupported_args("position", "occurrence")
1012        def regexpextract_sql(self, expression: exp.RegexpExtract) -> str:
1013            group = expression.args.get("group")
1014            params = expression.args.get("parameters")
1015
1016            # Do not render group if there is no following argument,
1017            # and it's the default value for this dialect
1018            if (
1019                not params
1020                and group
1021                and group.name == str(self.dialect.REGEXP_EXTRACT_DEFAULT_GROUP)
1022            ):
1023                group = None
1024            return self.func(
1025                "REGEXP_EXTRACT", expression.this, expression.expression, group, params
1026            )
@unsupported_args('culture')
def numbertostr_sql(self, expression: sqlglot.expressions.NumberToStr) -> str:
1028        @unsupported_args("culture")
1029        def numbertostr_sql(self, expression: exp.NumberToStr) -> str:
1030            fmt = expression.args.get("format")
1031            if fmt and fmt.is_int:
1032                return self.func("FORMAT", f"'{{:,.{fmt.name}f}}'", expression.this)
1033
1034            self.unsupported("Only integer formats are supported by NumberToStr")
1035            return self.function_fallback_sql(expression)
SELECT_KINDS: Tuple[str, ...] = ()
TRY_SUPPORTED = False
SUPPORTS_UESCAPE = False
AFTER_HAVING_MODIFIER_TRANSFORMS = {'windows': <function Generator.<lambda>>, 'qualify': <function Generator.<lambda>>}
Inherited Members
sqlglot.generator.Generator
Generator
NULL_ORDERING_SUPPORTED
LOCKING_READS_SUPPORTED
EXCEPT_INTERSECT_SUPPORT_ALL_CLAUSE
WRAP_DERIVED_VALUES
CREATE_FUNCTION_RETURN_AS
MATCHED_BY_SOURCE
SINGLE_STRING_INTERVAL
INTERVAL_ALLOWS_PLURAL_FORM
LIMIT_ONLY_LITERALS
GROUPINGS_SEP
INDEX_ON
QUERY_HINT_SEP
IS_BOOL_ALLOWED
DUPLICATE_KEY_UPDATE_WITH_SET
LIMIT_IS_TOP
RETURNING_END
EXTRACT_ALLOWS_QUOTES
TZ_TO_WITH_TIME_ZONE
VALUES_AS_TABLE
ALTER_TABLE_INCLUDE_COLUMN_KEYWORD
UNNEST_WITH_ORDINALITY
AGGREGATE_FILTER_SUPPORTED
COMPUTED_COLUMN_WITH_TYPE
SUPPORTS_TABLE_COPY
TABLESAMPLE_REQUIRES_PARENS
TABLESAMPLE_SIZE_IS_ROWS
TABLESAMPLE_WITH_METHOD
COLLATE_IS_FUNC
DATA_TYPE_SPECIFIERS_ALLOWED
ENSURE_BOOLS
CTE_RECURSIVE_KEYWORD_REQUIRED
SUPPORTS_SINGLE_ARG_CONCAT
SUPPORTS_TABLE_ALIAS_COLUMNS
UNPIVOT_ALIASES_ARE_IDENTIFIERS
INSERT_OVERWRITE
SUPPORTS_SELECT_INTO
SUPPORTS_UNLOGGED_TABLES
LIKE_PROPERTY_INSIDE_SCHEMA
JSON_TYPE_REQUIRED_FOR_EXTRACTION
JSON_PATH_SINGLE_QUOTE_ESCAPE
SET_OP_MODIFIERS
COPY_PARAMS_ARE_WRAPPED
COPY_PARAMS_EQ_REQUIRED
HEX_FUNC
WITH_PROPERTIES_PREFIX
QUOTE_JSON_PATH
SUPPORTS_EXPLODING_PROJECTIONS
SUPPORTS_CONVERT_TIMEZONE
SUPPORTS_MEDIAN
SUPPORTS_UNIX_SECONDS
PARSE_JSON_NAME
ARRAY_SIZE_NAME
TIME_PART_SINGULARS
TOKEN_MAPPING
EXPRESSION_PRECEDES_PROPERTIES_CREATABLES
WITH_SEPARATED_COMMENTS
EXCLUDE_COMMENTS
PARAMETERIZABLE_TEXT_TYPES
EXPRESSIONS_WITHOUT_NESTED_CTES
SENTINEL_LINE_BREAK
pretty
identify
normalize
pad
unsupported_level
max_unsupported
leading_comma
max_text_width
comments
dialect
normalize_functions
unsupported_messages
generate
preprocess
unsupported
sep
seg
pad_comment
maybe_comment
wrap
no_identify
normalize_func
indent
sql
uncache_sql
cache_sql
characterset_sql
column_parts
column_sql
columnposition_sql
columnconstraint_sql
computedcolumnconstraint_sql
autoincrementcolumnconstraint_sql
compresscolumnconstraint_sql
generatedasidentitycolumnconstraint_sql
generatedasrowcolumnconstraint_sql
periodforsystemtimeconstraint_sql
notnullcolumnconstraint_sql
transformcolumnconstraint_sql
primarykeycolumnconstraint_sql
uniquecolumnconstraint_sql
createable_sql
create_sql
sequenceproperties_sql
clone_sql
describe_sql
heredoc_sql
prepend_ctes
with_sql
cte_sql
tablealias_sql
bitstring_sql
hexstring_sql
bytestring_sql
unicodestring_sql
rawstring_sql
datatypeparam_sql
datatype_sql
directory_sql
delete_sql
drop_sql
set_operation
set_operations
fetch_sql
filter_sql
hint_sql
indexparameters_sql
index_sql
identifier_sql
hex_sql
lowerhex_sql
inputoutputformat_sql
national_sql
partition_sql
properties_sql
root_properties
properties
with_properties
locate_properties
property_name
property_sql
likeproperty_sql
fallbackproperty_sql
journalproperty_sql
freespaceproperty_sql
checksumproperty_sql
mergeblockratioproperty_sql
datablocksizeproperty_sql
blockcompressionproperty_sql
isolatedloadingproperty_sql
partitionboundspec_sql
partitionedofproperty_sql
lockingproperty_sql
withdataproperty_sql
withsystemversioningproperty_sql
insert_sql
introducer_sql
kill_sql
pseudotype_sql
objectidentifier_sql
onconflict_sql
returning_sql
rowformatdelimitedproperty_sql
withtablehint_sql
indextablehint_sql
historicaldata_sql
table_parts
table_sql
pivot_sql
version_sql
tuple_sql
update_sql
values_sql
var_sql
into_sql
from_sql
groupingsets_sql
rollup_sql
cube_sql
group_sql
having_sql
connect_sql
prior_sql
lambda_sql
lateral_op
lateral_sql
limit_sql
offset_sql
setitem_sql
set_sql
pragma_sql
lock_sql
literal_sql
escape_str
loaddata_sql
null_sql
boolean_sql
order_sql
withfill_sql
cluster_sql
distribute_sql
sort_sql
ordered_sql
matchrecognizemeasure_sql
matchrecognize_sql
query_modifiers
options_modifier
queryoption_sql
offset_limit_modifiers
after_limit_modifiers
select_sql
schema_sql
schema_columns_sql
star_sql
parameter_sql
sessionparameter_sql
placeholder_sql
subquery_sql
qualify_sql
prewhere_sql
where_sql
window_sql
partition_by_sql
windowspec_sql
between_sql
bracket_offset_expressions
all_sql
any_sql
exists_sql
case_sql
constraint_sql
nextvaluefor_sql
extract_sql
trim_sql
convert_concat_args
concat_sql
concatws_sql
check_sql
foreignkey_sql
primarykey_sql
if_sql
matchagainst_sql
jsonkeyvalue_sql
jsonpath_sql
json_path_part
formatjson_sql
jsonobject_sql
jsonobjectagg_sql
jsonarray_sql
jsonarrayagg_sql
jsoncolumndef_sql
jsonschema_sql
jsontable_sql
openjsoncolumndef_sql
openjson_sql
in_sql
in_unnest_op
return_sql
reference_sql
anonymous_sql
paren_sql
neg_sql
not_sql
alias_sql
pivotalias_sql
aliases_sql
atindex_sql
attimezone_sql
fromtimezone_sql
add_sql
and_sql
or_sql
xor_sql
connector_sql
bitwiseand_sql
bitwiseleftshift_sql
bitwisenot_sql
bitwiseor_sql
bitwiserightshift_sql
bitwisexor_sql
cast_sql
currentdate_sql
collate_sql
command_sql
comment_sql
mergetreettlaction_sql
mergetreettl_sql
transaction_sql
commit_sql
rollback_sql
altercolumn_sql
alterdiststyle_sql
altersortkey_sql
alterrename_sql
renamecolumn_sql
alterset_sql
alter_sql
add_column_sql
droppartition_sql
addconstraint_sql
distinct_sql
respectnulls_sql
havingmax_sql
intdiv_sql
dpipe_sql
div_sql
safedivide_sql
overlaps_sql
distance_sql
dot_sql
eq_sql
propertyeq_sql
escape_sql
glob_sql
gt_sql
gte_sql
ilike_sql
ilikeany_sql
is_sql
like_sql
likeany_sql
similarto_sql
lt_sql
lte_sql
mod_sql
mul_sql
neq_sql
nullsafeeq_sql
nullsafeneq_sql
slice_sql
sub_sql
trycast_sql
try_sql
log_sql
use_sql
binary
ceil_floor
function_fallback_sql
func
format_args
too_wide
format_time
expressions
op_expressions
naked_property
tag_sql
token_sql
userdefinedfunction_sql
joinhint_sql
kwarg_sql
when_sql
whens_sql
merge_sql
tochar_sql
tonumber_sql
dictproperty_sql
dictrange_sql
dictsubproperty_sql
duplicatekeyproperty_sql
uniquekeyproperty_sql
distributedbyproperty_sql
oncluster_sql
clusteredbyproperty_sql
anyvalue_sql
querytransform_sql
indexconstraintoption_sql
checkcolumnconstraint_sql
indexcolumnconstraint_sql
nvl2_sql
comprehension_sql
columnprefix_sql
opclass_sql
predict_sql
forin_sql
refresh_sql
toarray_sql
tsordstotime_sql
tsordstotimestamp_sql
tsordstodatetime_sql
tsordstodate_sql
unixdate_sql
lastday_sql
dateadd_sql
arrayany_sql
struct_sql
partitionrange_sql
truncatetable_sql
convert_sql
copyparameter_sql
credentials_sql
copy_sql
semicolon_sql
datadeletionproperty_sql
maskingpolicycolumnconstraint_sql
gapfill_sql
scope_resolution
scoperesolution_sql
rand_sql
changes_sql
pad_sql
summarize_sql
explodinggenerateseries_sql
arrayconcat_sql
converttimezone_sql
json_sql
jsonvalue_sql
conditionalinsert_sql
multitableinserts_sql
oncondition_sql
jsonextractquote_sql
jsonexists_sql
arrayagg_sql
apply_sql
grant_sql
grantprivilege_sql
grantprincipal_sql
columns_sql
overlay_sql
todouble_sql
string_sql
median_sql
overflowtruncatebehavior_sql
unixseconds_sql
arraysize_sql
attach_sql
detach_sql
attachoption_sql
featuresattime_sql
watermarkcolumnconstraint_sql
encodeproperty_sql
includeproperty_sql
xmlelement_sql
partitionbyrangeproperty_sql
partitionbyrangepropertydynamic_sql
unpivotcolumns_sql
analyzesample_sql
analyzestatistics_sql
analyzehistogram_sql
analyzedelete_sql
analyzelistchainedrows_sql
analyzevalidate_sql
analyze_sql
xmltable_sql
xmlnamespace_sql