sqlglot.tokens

   1from __future__ import annotations
   2
   3import os
   4import typing as t
   5from enum import auto
   6
   7from sqlglot.errors import SqlglotError, TokenError
   8from sqlglot.helper import AutoName
   9from sqlglot.trie import TrieResult, in_trie, new_trie
  10
  11if t.TYPE_CHECKING:
  12    from sqlglot.dialects.dialect import DialectType
  13
  14
  15try:
  16    from sqlglotrs import (  # type: ignore
  17        Tokenizer as RsTokenizer,
  18        TokenizerDialectSettings as RsTokenizerDialectSettings,
  19        TokenizerSettings as RsTokenizerSettings,
  20        TokenTypeSettings as RsTokenTypeSettings,
  21    )
  22
  23    USE_RS_TOKENIZER = os.environ.get("SQLGLOTRS_TOKENIZER", "1") == "1"
  24except ImportError:
  25    USE_RS_TOKENIZER = False
  26
  27
  28class TokenType(AutoName):
  29    L_PAREN = auto()
  30    R_PAREN = auto()
  31    L_BRACKET = auto()
  32    R_BRACKET = auto()
  33    L_BRACE = auto()
  34    R_BRACE = auto()
  35    COMMA = auto()
  36    DOT = auto()
  37    DASH = auto()
  38    PLUS = auto()
  39    COLON = auto()
  40    DOTCOLON = auto()
  41    DCOLON = auto()
  42    DQMARK = auto()
  43    SEMICOLON = auto()
  44    STAR = auto()
  45    BACKSLASH = auto()
  46    SLASH = auto()
  47    LT = auto()
  48    LTE = auto()
  49    GT = auto()
  50    GTE = auto()
  51    NOT = auto()
  52    EQ = auto()
  53    NEQ = auto()
  54    NULLSAFE_EQ = auto()
  55    COLON_EQ = auto()
  56    AND = auto()
  57    OR = auto()
  58    AMP = auto()
  59    DPIPE = auto()
  60    PIPE = auto()
  61    PIPE_SLASH = auto()
  62    DPIPE_SLASH = auto()
  63    CARET = auto()
  64    CARET_AT = auto()
  65    TILDA = auto()
  66    ARROW = auto()
  67    DARROW = auto()
  68    FARROW = auto()
  69    HASH = auto()
  70    HASH_ARROW = auto()
  71    DHASH_ARROW = auto()
  72    LR_ARROW = auto()
  73    DAT = auto()
  74    LT_AT = auto()
  75    AT_GT = auto()
  76    DOLLAR = auto()
  77    PARAMETER = auto()
  78    SESSION_PARAMETER = auto()
  79    DAMP = auto()
  80    XOR = auto()
  81    DSTAR = auto()
  82
  83    URI_START = auto()
  84
  85    BLOCK_START = auto()
  86    BLOCK_END = auto()
  87
  88    SPACE = auto()
  89    BREAK = auto()
  90
  91    STRING = auto()
  92    NUMBER = auto()
  93    IDENTIFIER = auto()
  94    DATABASE = auto()
  95    COLUMN = auto()
  96    COLUMN_DEF = auto()
  97    SCHEMA = auto()
  98    TABLE = auto()
  99    WAREHOUSE = auto()
 100    STAGE = auto()
 101    STREAMLIT = auto()
 102    VAR = auto()
 103    BIT_STRING = auto()
 104    HEX_STRING = auto()
 105    BYTE_STRING = auto()
 106    NATIONAL_STRING = auto()
 107    RAW_STRING = auto()
 108    HEREDOC_STRING = auto()
 109    UNICODE_STRING = auto()
 110
 111    # types
 112    BIT = auto()
 113    BOOLEAN = auto()
 114    TINYINT = auto()
 115    UTINYINT = auto()
 116    SMALLINT = auto()
 117    USMALLINT = auto()
 118    MEDIUMINT = auto()
 119    UMEDIUMINT = auto()
 120    INT = auto()
 121    UINT = auto()
 122    BIGINT = auto()
 123    UBIGINT = auto()
 124    INT128 = auto()
 125    UINT128 = auto()
 126    INT256 = auto()
 127    UINT256 = auto()
 128    FLOAT = auto()
 129    DOUBLE = auto()
 130    UDOUBLE = auto()
 131    DECIMAL = auto()
 132    DECIMAL32 = auto()
 133    DECIMAL64 = auto()
 134    DECIMAL128 = auto()
 135    DECIMAL256 = auto()
 136    UDECIMAL = auto()
 137    BIGDECIMAL = auto()
 138    CHAR = auto()
 139    NCHAR = auto()
 140    VARCHAR = auto()
 141    NVARCHAR = auto()
 142    BPCHAR = auto()
 143    TEXT = auto()
 144    MEDIUMTEXT = auto()
 145    LONGTEXT = auto()
 146    BLOB = auto()
 147    MEDIUMBLOB = auto()
 148    LONGBLOB = auto()
 149    TINYBLOB = auto()
 150    TINYTEXT = auto()
 151    NAME = auto()
 152    BINARY = auto()
 153    VARBINARY = auto()
 154    JSON = auto()
 155    JSONB = auto()
 156    TIME = auto()
 157    TIMETZ = auto()
 158    TIMESTAMP = auto()
 159    TIMESTAMPTZ = auto()
 160    TIMESTAMPLTZ = auto()
 161    TIMESTAMPNTZ = auto()
 162    TIMESTAMP_S = auto()
 163    TIMESTAMP_MS = auto()
 164    TIMESTAMP_NS = auto()
 165    DATETIME = auto()
 166    DATETIME2 = auto()
 167    DATETIME64 = auto()
 168    SMALLDATETIME = auto()
 169    DATE = auto()
 170    DATE32 = auto()
 171    INT4RANGE = auto()
 172    INT4MULTIRANGE = auto()
 173    INT8RANGE = auto()
 174    INT8MULTIRANGE = auto()
 175    NUMRANGE = auto()
 176    NUMMULTIRANGE = auto()
 177    TSRANGE = auto()
 178    TSMULTIRANGE = auto()
 179    TSTZRANGE = auto()
 180    TSTZMULTIRANGE = auto()
 181    DATERANGE = auto()
 182    DATEMULTIRANGE = auto()
 183    UUID = auto()
 184    GEOGRAPHY = auto()
 185    NULLABLE = auto()
 186    GEOMETRY = auto()
 187    POINT = auto()
 188    RING = auto()
 189    LINESTRING = auto()
 190    MULTILINESTRING = auto()
 191    POLYGON = auto()
 192    MULTIPOLYGON = auto()
 193    HLLSKETCH = auto()
 194    HSTORE = auto()
 195    SUPER = auto()
 196    SERIAL = auto()
 197    SMALLSERIAL = auto()
 198    BIGSERIAL = auto()
 199    XML = auto()
 200    YEAR = auto()
 201    USERDEFINED = auto()
 202    MONEY = auto()
 203    SMALLMONEY = auto()
 204    ROWVERSION = auto()
 205    IMAGE = auto()
 206    VARIANT = auto()
 207    OBJECT = auto()
 208    INET = auto()
 209    IPADDRESS = auto()
 210    IPPREFIX = auto()
 211    IPV4 = auto()
 212    IPV6 = auto()
 213    ENUM = auto()
 214    ENUM8 = auto()
 215    ENUM16 = auto()
 216    FIXEDSTRING = auto()
 217    LOWCARDINALITY = auto()
 218    NESTED = auto()
 219    AGGREGATEFUNCTION = auto()
 220    SIMPLEAGGREGATEFUNCTION = auto()
 221    TDIGEST = auto()
 222    UNKNOWN = auto()
 223    VECTOR = auto()
 224    DYNAMIC = auto()
 225
 226    # keywords
 227    ALIAS = auto()
 228    ALTER = auto()
 229    ALWAYS = auto()
 230    ALL = auto()
 231    ANTI = auto()
 232    ANY = auto()
 233    APPLY = auto()
 234    ARRAY = auto()
 235    ASC = auto()
 236    ASOF = auto()
 237    ATTACH = auto()
 238    AUTO_INCREMENT = auto()
 239    BEGIN = auto()
 240    BETWEEN = auto()
 241    BULK_COLLECT_INTO = auto()
 242    CACHE = auto()
 243    CASE = auto()
 244    CHARACTER_SET = auto()
 245    CLUSTER_BY = auto()
 246    COLLATE = auto()
 247    COMMAND = auto()
 248    COMMENT = auto()
 249    COMMIT = auto()
 250    CONNECT_BY = auto()
 251    CONSTRAINT = auto()
 252    COPY = auto()
 253    CREATE = auto()
 254    CROSS = auto()
 255    CUBE = auto()
 256    CURRENT_DATE = auto()
 257    CURRENT_DATETIME = auto()
 258    CURRENT_SCHEMA = auto()
 259    CURRENT_TIME = auto()
 260    CURRENT_TIMESTAMP = auto()
 261    CURRENT_USER = auto()
 262    DECLARE = auto()
 263    DEFAULT = auto()
 264    DELETE = auto()
 265    DESC = auto()
 266    DESCRIBE = auto()
 267    DETACH = auto()
 268    DICTIONARY = auto()
 269    DISTINCT = auto()
 270    DISTRIBUTE_BY = auto()
 271    DIV = auto()
 272    DROP = auto()
 273    ELSE = auto()
 274    END = auto()
 275    ESCAPE = auto()
 276    EXCEPT = auto()
 277    EXECUTE = auto()
 278    EXISTS = auto()
 279    FALSE = auto()
 280    FETCH = auto()
 281    FILE_FORMAT = auto()
 282    FILTER = auto()
 283    FINAL = auto()
 284    FIRST = auto()
 285    FOR = auto()
 286    FORCE = auto()
 287    FOREIGN_KEY = auto()
 288    FORMAT = auto()
 289    FROM = auto()
 290    FULL = auto()
 291    FUNCTION = auto()
 292    GLOB = auto()
 293    GLOBAL = auto()
 294    GRANT = auto()
 295    GROUP_BY = auto()
 296    GROUPING_SETS = auto()
 297    HAVING = auto()
 298    HINT = auto()
 299    IGNORE = auto()
 300    ILIKE = auto()
 301    ILIKE_ANY = auto()
 302    IN = auto()
 303    INDEX = auto()
 304    INNER = auto()
 305    INSERT = auto()
 306    INTERSECT = auto()
 307    INTERVAL = auto()
 308    INTO = auto()
 309    INTRODUCER = auto()
 310    IRLIKE = auto()
 311    IS = auto()
 312    ISNULL = auto()
 313    JOIN = auto()
 314    JOIN_MARKER = auto()
 315    KEEP = auto()
 316    KEY = auto()
 317    KILL = auto()
 318    LANGUAGE = auto()
 319    LATERAL = auto()
 320    LEFT = auto()
 321    LIKE = auto()
 322    LIKE_ANY = auto()
 323    LIMIT = auto()
 324    LIST = auto()
 325    LOAD = auto()
 326    LOCK = auto()
 327    MAP = auto()
 328    MATCH_CONDITION = auto()
 329    MATCH_RECOGNIZE = auto()
 330    MEMBER_OF = auto()
 331    MERGE = auto()
 332    MOD = auto()
 333    MODEL = auto()
 334    NATURAL = auto()
 335    NEXT = auto()
 336    NOTNULL = auto()
 337    NULL = auto()
 338    OBJECT_IDENTIFIER = auto()
 339    OFFSET = auto()
 340    ON = auto()
 341    ONLY = auto()
 342    OPERATOR = auto()
 343    ORDER_BY = auto()
 344    ORDER_SIBLINGS_BY = auto()
 345    ORDERED = auto()
 346    ORDINALITY = auto()
 347    OUTER = auto()
 348    OVER = auto()
 349    OVERLAPS = auto()
 350    OVERWRITE = auto()
 351    PARTITION = auto()
 352    PARTITION_BY = auto()
 353    PERCENT = auto()
 354    PIVOT = auto()
 355    PLACEHOLDER = auto()
 356    POSITIONAL = auto()
 357    PRAGMA = auto()
 358    PREWHERE = auto()
 359    PRIMARY_KEY = auto()
 360    PROCEDURE = auto()
 361    PROPERTIES = auto()
 362    PSEUDO_TYPE = auto()
 363    PUT = auto()
 364    QUALIFY = auto()
 365    QUOTE = auto()
 366    RANGE = auto()
 367    RECURSIVE = auto()
 368    REFRESH = auto()
 369    RENAME = auto()
 370    REPLACE = auto()
 371    RETURNING = auto()
 372    REFERENCES = auto()
 373    RIGHT = auto()
 374    RLIKE = auto()
 375    ROLLBACK = auto()
 376    ROLLUP = auto()
 377    ROW = auto()
 378    ROWS = auto()
 379    SELECT = auto()
 380    SEMI = auto()
 381    SEPARATOR = auto()
 382    SEQUENCE = auto()
 383    SERDE_PROPERTIES = auto()
 384    SET = auto()
 385    SETTINGS = auto()
 386    SHOW = auto()
 387    SIMILAR_TO = auto()
 388    SOME = auto()
 389    SORT_BY = auto()
 390    START_WITH = auto()
 391    STORAGE_INTEGRATION = auto()
 392    STRAIGHT_JOIN = auto()
 393    STRUCT = auto()
 394    SUMMARIZE = auto()
 395    TABLE_SAMPLE = auto()
 396    TAG = auto()
 397    TEMPORARY = auto()
 398    TOP = auto()
 399    THEN = auto()
 400    TRUE = auto()
 401    TRUNCATE = auto()
 402    UNCACHE = auto()
 403    UNION = auto()
 404    UNNEST = auto()
 405    UNPIVOT = auto()
 406    UPDATE = auto()
 407    USE = auto()
 408    USING = auto()
 409    VALUES = auto()
 410    VIEW = auto()
 411    VOLATILE = auto()
 412    WHEN = auto()
 413    WHERE = auto()
 414    WINDOW = auto()
 415    WITH = auto()
 416    UNIQUE = auto()
 417    VERSION_SNAPSHOT = auto()
 418    TIMESTAMP_SNAPSHOT = auto()
 419    OPTION = auto()
 420    SINK = auto()
 421    SOURCE = auto()
 422    ANALYZE = auto()
 423    NAMESPACE = auto()
 424    EXPORT = auto()
 425
 426
 427_ALL_TOKEN_TYPES = list(TokenType)
 428_TOKEN_TYPE_TO_INDEX = {token_type: i for i, token_type in enumerate(_ALL_TOKEN_TYPES)}
 429
 430
 431class Token:
 432    __slots__ = ("token_type", "text", "line", "col", "start", "end", "comments")
 433
 434    @classmethod
 435    def number(cls, number: int) -> Token:
 436        """Returns a NUMBER token with `number` as its text."""
 437        return cls(TokenType.NUMBER, str(number))
 438
 439    @classmethod
 440    def string(cls, string: str) -> Token:
 441        """Returns a STRING token with `string` as its text."""
 442        return cls(TokenType.STRING, string)
 443
 444    @classmethod
 445    def identifier(cls, identifier: str) -> Token:
 446        """Returns an IDENTIFIER token with `identifier` as its text."""
 447        return cls(TokenType.IDENTIFIER, identifier)
 448
 449    @classmethod
 450    def var(cls, var: str) -> Token:
 451        """Returns an VAR token with `var` as its text."""
 452        return cls(TokenType.VAR, var)
 453
 454    def __init__(
 455        self,
 456        token_type: TokenType,
 457        text: str,
 458        line: int = 1,
 459        col: int = 1,
 460        start: int = 0,
 461        end: int = 0,
 462        comments: t.Optional[t.List[str]] = None,
 463    ) -> None:
 464        """Token initializer.
 465
 466        Args:
 467            token_type: The TokenType Enum.
 468            text: The text of the token.
 469            line: The line that the token ends on.
 470            col: The column that the token ends on.
 471            start: The start index of the token.
 472            end: The ending index of the token.
 473            comments: The comments to attach to the token.
 474        """
 475        self.token_type = token_type
 476        self.text = text
 477        self.line = line
 478        self.col = col
 479        self.start = start
 480        self.end = end
 481        self.comments = [] if comments is None else comments
 482
 483    def __repr__(self) -> str:
 484        attributes = ", ".join(f"{k}: {getattr(self, k)}" for k in self.__slots__)
 485        return f"<Token {attributes}>"
 486
 487
 488class _Tokenizer(type):
 489    def __new__(cls, clsname, bases, attrs):
 490        klass = super().__new__(cls, clsname, bases, attrs)
 491
 492        def _convert_quotes(arr: t.List[str | t.Tuple[str, str]]) -> t.Dict[str, str]:
 493            return dict(
 494                (item, item) if isinstance(item, str) else (item[0], item[1]) for item in arr
 495            )
 496
 497        def _quotes_to_format(
 498            token_type: TokenType, arr: t.List[str | t.Tuple[str, str]]
 499        ) -> t.Dict[str, t.Tuple[str, TokenType]]:
 500            return {k: (v, token_type) for k, v in _convert_quotes(arr).items()}
 501
 502        klass._QUOTES = _convert_quotes(klass.QUOTES)
 503        klass._IDENTIFIERS = _convert_quotes(klass.IDENTIFIERS)
 504
 505        klass._FORMAT_STRINGS = {
 506            **{
 507                p + s: (e, TokenType.NATIONAL_STRING)
 508                for s, e in klass._QUOTES.items()
 509                for p in ("n", "N")
 510            },
 511            **_quotes_to_format(TokenType.BIT_STRING, klass.BIT_STRINGS),
 512            **_quotes_to_format(TokenType.BYTE_STRING, klass.BYTE_STRINGS),
 513            **_quotes_to_format(TokenType.HEX_STRING, klass.HEX_STRINGS),
 514            **_quotes_to_format(TokenType.RAW_STRING, klass.RAW_STRINGS),
 515            **_quotes_to_format(TokenType.HEREDOC_STRING, klass.HEREDOC_STRINGS),
 516            **_quotes_to_format(TokenType.UNICODE_STRING, klass.UNICODE_STRINGS),
 517        }
 518
 519        klass._STRING_ESCAPES = set(klass.STRING_ESCAPES)
 520        klass._IDENTIFIER_ESCAPES = set(klass.IDENTIFIER_ESCAPES)
 521        klass._COMMENTS = {
 522            **dict(
 523                (comment, None) if isinstance(comment, str) else (comment[0], comment[1])
 524                for comment in klass.COMMENTS
 525            ),
 526            "{#": "#}",  # Ensure Jinja comments are tokenized correctly in all dialects
 527        }
 528        if klass.HINT_START in klass.KEYWORDS:
 529            klass._COMMENTS[klass.HINT_START] = "*/"
 530
 531        klass._KEYWORD_TRIE = new_trie(
 532            key.upper()
 533            for key in (
 534                *klass.KEYWORDS,
 535                *klass._COMMENTS,
 536                *klass._QUOTES,
 537                *klass._FORMAT_STRINGS,
 538            )
 539            if " " in key or any(single in key for single in klass.SINGLE_TOKENS)
 540        )
 541
 542        if USE_RS_TOKENIZER:
 543            settings = RsTokenizerSettings(
 544                white_space={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.WHITE_SPACE.items()},
 545                single_tokens={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.SINGLE_TOKENS.items()},
 546                keywords={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.KEYWORDS.items()},
 547                numeric_literals=klass.NUMERIC_LITERALS,
 548                identifiers=klass._IDENTIFIERS,
 549                identifier_escapes=klass._IDENTIFIER_ESCAPES,
 550                string_escapes=klass._STRING_ESCAPES,
 551                quotes=klass._QUOTES,
 552                format_strings={
 553                    k: (v1, _TOKEN_TYPE_TO_INDEX[v2])
 554                    for k, (v1, v2) in klass._FORMAT_STRINGS.items()
 555                },
 556                has_bit_strings=bool(klass.BIT_STRINGS),
 557                has_hex_strings=bool(klass.HEX_STRINGS),
 558                comments=klass._COMMENTS,
 559                var_single_tokens=klass.VAR_SINGLE_TOKENS,
 560                commands={_TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMANDS},
 561                command_prefix_tokens={
 562                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMAND_PREFIX_TOKENS
 563                },
 564                heredoc_tag_is_identifier=klass.HEREDOC_TAG_IS_IDENTIFIER,
 565                string_escapes_allowed_in_raw_strings=klass.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS,
 566                nested_comments=klass.NESTED_COMMENTS,
 567                hint_start=klass.HINT_START,
 568                tokens_preceding_hint={
 569                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.TOKENS_PRECEDING_HINT
 570                },
 571            )
 572            token_types = RsTokenTypeSettings(
 573                bit_string=_TOKEN_TYPE_TO_INDEX[TokenType.BIT_STRING],
 574                break_=_TOKEN_TYPE_TO_INDEX[TokenType.BREAK],
 575                dcolon=_TOKEN_TYPE_TO_INDEX[TokenType.DCOLON],
 576                heredoc_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEREDOC_STRING],
 577                raw_string=_TOKEN_TYPE_TO_INDEX[TokenType.RAW_STRING],
 578                hex_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEX_STRING],
 579                identifier=_TOKEN_TYPE_TO_INDEX[TokenType.IDENTIFIER],
 580                number=_TOKEN_TYPE_TO_INDEX[TokenType.NUMBER],
 581                parameter=_TOKEN_TYPE_TO_INDEX[TokenType.PARAMETER],
 582                semicolon=_TOKEN_TYPE_TO_INDEX[TokenType.SEMICOLON],
 583                string=_TOKEN_TYPE_TO_INDEX[TokenType.STRING],
 584                var=_TOKEN_TYPE_TO_INDEX[TokenType.VAR],
 585                heredoc_string_alternative=_TOKEN_TYPE_TO_INDEX[klass.HEREDOC_STRING_ALTERNATIVE],
 586                hint=_TOKEN_TYPE_TO_INDEX[TokenType.HINT],
 587            )
 588            klass._RS_TOKENIZER = RsTokenizer(settings, token_types)
 589        else:
 590            klass._RS_TOKENIZER = None
 591
 592        return klass
 593
 594
 595class Tokenizer(metaclass=_Tokenizer):
 596    SINGLE_TOKENS = {
 597        "(": TokenType.L_PAREN,
 598        ")": TokenType.R_PAREN,
 599        "[": TokenType.L_BRACKET,
 600        "]": TokenType.R_BRACKET,
 601        "{": TokenType.L_BRACE,
 602        "}": TokenType.R_BRACE,
 603        "&": TokenType.AMP,
 604        "^": TokenType.CARET,
 605        ":": TokenType.COLON,
 606        ",": TokenType.COMMA,
 607        ".": TokenType.DOT,
 608        "-": TokenType.DASH,
 609        "=": TokenType.EQ,
 610        ">": TokenType.GT,
 611        "<": TokenType.LT,
 612        "%": TokenType.MOD,
 613        "!": TokenType.NOT,
 614        "|": TokenType.PIPE,
 615        "+": TokenType.PLUS,
 616        ";": TokenType.SEMICOLON,
 617        "/": TokenType.SLASH,
 618        "\\": TokenType.BACKSLASH,
 619        "*": TokenType.STAR,
 620        "~": TokenType.TILDA,
 621        "?": TokenType.PLACEHOLDER,
 622        "@": TokenType.PARAMETER,
 623        "#": TokenType.HASH,
 624        # Used for breaking a var like x'y' but nothing else the token type doesn't matter
 625        "'": TokenType.UNKNOWN,
 626        "`": TokenType.UNKNOWN,
 627        '"': TokenType.UNKNOWN,
 628    }
 629
 630    BIT_STRINGS: t.List[str | t.Tuple[str, str]] = []
 631    BYTE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 632    HEX_STRINGS: t.List[str | t.Tuple[str, str]] = []
 633    RAW_STRINGS: t.List[str | t.Tuple[str, str]] = []
 634    HEREDOC_STRINGS: t.List[str | t.Tuple[str, str]] = []
 635    UNICODE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 636    IDENTIFIERS: t.List[str | t.Tuple[str, str]] = ['"']
 637    QUOTES: t.List[t.Tuple[str, str] | str] = ["'"]
 638    STRING_ESCAPES = ["'"]
 639    VAR_SINGLE_TOKENS: t.Set[str] = set()
 640
 641    # The strings in this list can always be used as escapes, regardless of the surrounding
 642    # identifier delimiters. By default, the closing delimiter is assumed to also act as an
 643    # identifier escape, e.g. if we use double-quotes, then they also act as escapes: "x"""
 644    IDENTIFIER_ESCAPES: t.List[str] = []
 645
 646    # Whether the heredoc tags follow the same lexical rules as unquoted identifiers
 647    HEREDOC_TAG_IS_IDENTIFIER = False
 648
 649    # Token that we'll generate as a fallback if the heredoc prefix doesn't correspond to a heredoc
 650    HEREDOC_STRING_ALTERNATIVE = TokenType.VAR
 651
 652    # Whether string escape characters function as such when placed within raw strings
 653    STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS = True
 654
 655    NESTED_COMMENTS = True
 656
 657    HINT_START = "/*+"
 658
 659    TOKENS_PRECEDING_HINT = {TokenType.SELECT, TokenType.INSERT, TokenType.UPDATE, TokenType.DELETE}
 660
 661    # Autofilled
 662    _COMMENTS: t.Dict[str, str] = {}
 663    _FORMAT_STRINGS: t.Dict[str, t.Tuple[str, TokenType]] = {}
 664    _IDENTIFIERS: t.Dict[str, str] = {}
 665    _IDENTIFIER_ESCAPES: t.Set[str] = set()
 666    _QUOTES: t.Dict[str, str] = {}
 667    _STRING_ESCAPES: t.Set[str] = set()
 668    _KEYWORD_TRIE: t.Dict = {}
 669    _RS_TOKENIZER: t.Optional[t.Any] = None
 670
 671    KEYWORDS: t.Dict[str, TokenType] = {
 672        **{f"{{%{postfix}": TokenType.BLOCK_START for postfix in ("", "+", "-")},
 673        **{f"{prefix}%}}": TokenType.BLOCK_END for prefix in ("", "+", "-")},
 674        **{f"{{{{{postfix}": TokenType.BLOCK_START for postfix in ("+", "-")},
 675        **{f"{prefix}}}}}": TokenType.BLOCK_END for prefix in ("+", "-")},
 676        HINT_START: TokenType.HINT,
 677        "==": TokenType.EQ,
 678        "::": TokenType.DCOLON,
 679        "||": TokenType.DPIPE,
 680        ">=": TokenType.GTE,
 681        "<=": TokenType.LTE,
 682        "<>": TokenType.NEQ,
 683        "!=": TokenType.NEQ,
 684        ":=": TokenType.COLON_EQ,
 685        "<=>": TokenType.NULLSAFE_EQ,
 686        "->": TokenType.ARROW,
 687        "->>": TokenType.DARROW,
 688        "=>": TokenType.FARROW,
 689        "#>": TokenType.HASH_ARROW,
 690        "#>>": TokenType.DHASH_ARROW,
 691        "<->": TokenType.LR_ARROW,
 692        "&&": TokenType.DAMP,
 693        "??": TokenType.DQMARK,
 694        "~~~": TokenType.GLOB,
 695        "~~": TokenType.LIKE,
 696        "~~*": TokenType.ILIKE,
 697        "~*": TokenType.IRLIKE,
 698        "ALL": TokenType.ALL,
 699        "ALWAYS": TokenType.ALWAYS,
 700        "AND": TokenType.AND,
 701        "ANTI": TokenType.ANTI,
 702        "ANY": TokenType.ANY,
 703        "ASC": TokenType.ASC,
 704        "AS": TokenType.ALIAS,
 705        "ASOF": TokenType.ASOF,
 706        "AUTOINCREMENT": TokenType.AUTO_INCREMENT,
 707        "AUTO_INCREMENT": TokenType.AUTO_INCREMENT,
 708        "BEGIN": TokenType.BEGIN,
 709        "BETWEEN": TokenType.BETWEEN,
 710        "CACHE": TokenType.CACHE,
 711        "UNCACHE": TokenType.UNCACHE,
 712        "CASE": TokenType.CASE,
 713        "CHARACTER SET": TokenType.CHARACTER_SET,
 714        "CLUSTER BY": TokenType.CLUSTER_BY,
 715        "COLLATE": TokenType.COLLATE,
 716        "COLUMN": TokenType.COLUMN,
 717        "COMMIT": TokenType.COMMIT,
 718        "CONNECT BY": TokenType.CONNECT_BY,
 719        "CONSTRAINT": TokenType.CONSTRAINT,
 720        "COPY": TokenType.COPY,
 721        "CREATE": TokenType.CREATE,
 722        "CROSS": TokenType.CROSS,
 723        "CUBE": TokenType.CUBE,
 724        "CURRENT_DATE": TokenType.CURRENT_DATE,
 725        "CURRENT_SCHEMA": TokenType.CURRENT_SCHEMA,
 726        "CURRENT_TIME": TokenType.CURRENT_TIME,
 727        "CURRENT_TIMESTAMP": TokenType.CURRENT_TIMESTAMP,
 728        "CURRENT_USER": TokenType.CURRENT_USER,
 729        "DATABASE": TokenType.DATABASE,
 730        "DEFAULT": TokenType.DEFAULT,
 731        "DELETE": TokenType.DELETE,
 732        "DESC": TokenType.DESC,
 733        "DESCRIBE": TokenType.DESCRIBE,
 734        "DISTINCT": TokenType.DISTINCT,
 735        "DISTRIBUTE BY": TokenType.DISTRIBUTE_BY,
 736        "DIV": TokenType.DIV,
 737        "DROP": TokenType.DROP,
 738        "ELSE": TokenType.ELSE,
 739        "END": TokenType.END,
 740        "ENUM": TokenType.ENUM,
 741        "ESCAPE": TokenType.ESCAPE,
 742        "EXCEPT": TokenType.EXCEPT,
 743        "EXECUTE": TokenType.EXECUTE,
 744        "EXISTS": TokenType.EXISTS,
 745        "FALSE": TokenType.FALSE,
 746        "FETCH": TokenType.FETCH,
 747        "FILTER": TokenType.FILTER,
 748        "FIRST": TokenType.FIRST,
 749        "FULL": TokenType.FULL,
 750        "FUNCTION": TokenType.FUNCTION,
 751        "FOR": TokenType.FOR,
 752        "FOREIGN KEY": TokenType.FOREIGN_KEY,
 753        "FORMAT": TokenType.FORMAT,
 754        "FROM": TokenType.FROM,
 755        "GEOGRAPHY": TokenType.GEOGRAPHY,
 756        "GEOMETRY": TokenType.GEOMETRY,
 757        "GLOB": TokenType.GLOB,
 758        "GROUP BY": TokenType.GROUP_BY,
 759        "GROUPING SETS": TokenType.GROUPING_SETS,
 760        "HAVING": TokenType.HAVING,
 761        "ILIKE": TokenType.ILIKE,
 762        "IN": TokenType.IN,
 763        "INDEX": TokenType.INDEX,
 764        "INET": TokenType.INET,
 765        "INNER": TokenType.INNER,
 766        "INSERT": TokenType.INSERT,
 767        "INTERVAL": TokenType.INTERVAL,
 768        "INTERSECT": TokenType.INTERSECT,
 769        "INTO": TokenType.INTO,
 770        "IS": TokenType.IS,
 771        "ISNULL": TokenType.ISNULL,
 772        "JOIN": TokenType.JOIN,
 773        "KEEP": TokenType.KEEP,
 774        "KILL": TokenType.KILL,
 775        "LATERAL": TokenType.LATERAL,
 776        "LEFT": TokenType.LEFT,
 777        "LIKE": TokenType.LIKE,
 778        "LIMIT": TokenType.LIMIT,
 779        "LOAD": TokenType.LOAD,
 780        "LOCK": TokenType.LOCK,
 781        "MERGE": TokenType.MERGE,
 782        "NAMESPACE": TokenType.NAMESPACE,
 783        "NATURAL": TokenType.NATURAL,
 784        "NEXT": TokenType.NEXT,
 785        "NOT": TokenType.NOT,
 786        "NOTNULL": TokenType.NOTNULL,
 787        "NULL": TokenType.NULL,
 788        "OBJECT": TokenType.OBJECT,
 789        "OFFSET": TokenType.OFFSET,
 790        "ON": TokenType.ON,
 791        "OR": TokenType.OR,
 792        "XOR": TokenType.XOR,
 793        "ORDER BY": TokenType.ORDER_BY,
 794        "ORDINALITY": TokenType.ORDINALITY,
 795        "OUTER": TokenType.OUTER,
 796        "OVER": TokenType.OVER,
 797        "OVERLAPS": TokenType.OVERLAPS,
 798        "OVERWRITE": TokenType.OVERWRITE,
 799        "PARTITION": TokenType.PARTITION,
 800        "PARTITION BY": TokenType.PARTITION_BY,
 801        "PARTITIONED BY": TokenType.PARTITION_BY,
 802        "PARTITIONED_BY": TokenType.PARTITION_BY,
 803        "PERCENT": TokenType.PERCENT,
 804        "PIVOT": TokenType.PIVOT,
 805        "PRAGMA": TokenType.PRAGMA,
 806        "PRIMARY KEY": TokenType.PRIMARY_KEY,
 807        "PROCEDURE": TokenType.PROCEDURE,
 808        "QUALIFY": TokenType.QUALIFY,
 809        "RANGE": TokenType.RANGE,
 810        "RECURSIVE": TokenType.RECURSIVE,
 811        "REGEXP": TokenType.RLIKE,
 812        "RENAME": TokenType.RENAME,
 813        "REPLACE": TokenType.REPLACE,
 814        "RETURNING": TokenType.RETURNING,
 815        "REFERENCES": TokenType.REFERENCES,
 816        "RIGHT": TokenType.RIGHT,
 817        "RLIKE": TokenType.RLIKE,
 818        "ROLLBACK": TokenType.ROLLBACK,
 819        "ROLLUP": TokenType.ROLLUP,
 820        "ROW": TokenType.ROW,
 821        "ROWS": TokenType.ROWS,
 822        "SCHEMA": TokenType.SCHEMA,
 823        "SELECT": TokenType.SELECT,
 824        "SEMI": TokenType.SEMI,
 825        "SET": TokenType.SET,
 826        "SETTINGS": TokenType.SETTINGS,
 827        "SHOW": TokenType.SHOW,
 828        "SIMILAR TO": TokenType.SIMILAR_TO,
 829        "SOME": TokenType.SOME,
 830        "SORT BY": TokenType.SORT_BY,
 831        "START WITH": TokenType.START_WITH,
 832        "STRAIGHT_JOIN": TokenType.STRAIGHT_JOIN,
 833        "TABLE": TokenType.TABLE,
 834        "TABLESAMPLE": TokenType.TABLE_SAMPLE,
 835        "TEMP": TokenType.TEMPORARY,
 836        "TEMPORARY": TokenType.TEMPORARY,
 837        "THEN": TokenType.THEN,
 838        "TRUE": TokenType.TRUE,
 839        "TRUNCATE": TokenType.TRUNCATE,
 840        "UNION": TokenType.UNION,
 841        "UNKNOWN": TokenType.UNKNOWN,
 842        "UNNEST": TokenType.UNNEST,
 843        "UNPIVOT": TokenType.UNPIVOT,
 844        "UPDATE": TokenType.UPDATE,
 845        "USE": TokenType.USE,
 846        "USING": TokenType.USING,
 847        "UUID": TokenType.UUID,
 848        "VALUES": TokenType.VALUES,
 849        "VIEW": TokenType.VIEW,
 850        "VOLATILE": TokenType.VOLATILE,
 851        "WHEN": TokenType.WHEN,
 852        "WHERE": TokenType.WHERE,
 853        "WINDOW": TokenType.WINDOW,
 854        "WITH": TokenType.WITH,
 855        "APPLY": TokenType.APPLY,
 856        "ARRAY": TokenType.ARRAY,
 857        "BIT": TokenType.BIT,
 858        "BOOL": TokenType.BOOLEAN,
 859        "BOOLEAN": TokenType.BOOLEAN,
 860        "BYTE": TokenType.TINYINT,
 861        "MEDIUMINT": TokenType.MEDIUMINT,
 862        "INT1": TokenType.TINYINT,
 863        "TINYINT": TokenType.TINYINT,
 864        "INT16": TokenType.SMALLINT,
 865        "SHORT": TokenType.SMALLINT,
 866        "SMALLINT": TokenType.SMALLINT,
 867        "HUGEINT": TokenType.INT128,
 868        "UHUGEINT": TokenType.UINT128,
 869        "INT2": TokenType.SMALLINT,
 870        "INTEGER": TokenType.INT,
 871        "INT": TokenType.INT,
 872        "INT4": TokenType.INT,
 873        "INT32": TokenType.INT,
 874        "INT64": TokenType.BIGINT,
 875        "INT128": TokenType.INT128,
 876        "INT256": TokenType.INT256,
 877        "LONG": TokenType.BIGINT,
 878        "BIGINT": TokenType.BIGINT,
 879        "INT8": TokenType.TINYINT,
 880        "UINT": TokenType.UINT,
 881        "UINT128": TokenType.UINT128,
 882        "UINT256": TokenType.UINT256,
 883        "DEC": TokenType.DECIMAL,
 884        "DECIMAL": TokenType.DECIMAL,
 885        "DECIMAL32": TokenType.DECIMAL32,
 886        "DECIMAL64": TokenType.DECIMAL64,
 887        "DECIMAL128": TokenType.DECIMAL128,
 888        "DECIMAL256": TokenType.DECIMAL256,
 889        "BIGDECIMAL": TokenType.BIGDECIMAL,
 890        "BIGNUMERIC": TokenType.BIGDECIMAL,
 891        "LIST": TokenType.LIST,
 892        "MAP": TokenType.MAP,
 893        "NULLABLE": TokenType.NULLABLE,
 894        "NUMBER": TokenType.DECIMAL,
 895        "NUMERIC": TokenType.DECIMAL,
 896        "FIXED": TokenType.DECIMAL,
 897        "REAL": TokenType.FLOAT,
 898        "FLOAT": TokenType.FLOAT,
 899        "FLOAT4": TokenType.FLOAT,
 900        "FLOAT8": TokenType.DOUBLE,
 901        "DOUBLE": TokenType.DOUBLE,
 902        "DOUBLE PRECISION": TokenType.DOUBLE,
 903        "JSON": TokenType.JSON,
 904        "JSONB": TokenType.JSONB,
 905        "CHAR": TokenType.CHAR,
 906        "CHARACTER": TokenType.CHAR,
 907        "NCHAR": TokenType.NCHAR,
 908        "VARCHAR": TokenType.VARCHAR,
 909        "VARCHAR2": TokenType.VARCHAR,
 910        "NVARCHAR": TokenType.NVARCHAR,
 911        "NVARCHAR2": TokenType.NVARCHAR,
 912        "BPCHAR": TokenType.BPCHAR,
 913        "STR": TokenType.TEXT,
 914        "STRING": TokenType.TEXT,
 915        "TEXT": TokenType.TEXT,
 916        "LONGTEXT": TokenType.LONGTEXT,
 917        "MEDIUMTEXT": TokenType.MEDIUMTEXT,
 918        "TINYTEXT": TokenType.TINYTEXT,
 919        "CLOB": TokenType.TEXT,
 920        "LONGVARCHAR": TokenType.TEXT,
 921        "BINARY": TokenType.BINARY,
 922        "BLOB": TokenType.VARBINARY,
 923        "LONGBLOB": TokenType.LONGBLOB,
 924        "MEDIUMBLOB": TokenType.MEDIUMBLOB,
 925        "TINYBLOB": TokenType.TINYBLOB,
 926        "BYTEA": TokenType.VARBINARY,
 927        "VARBINARY": TokenType.VARBINARY,
 928        "TIME": TokenType.TIME,
 929        "TIMETZ": TokenType.TIMETZ,
 930        "TIMESTAMP": TokenType.TIMESTAMP,
 931        "TIMESTAMPTZ": TokenType.TIMESTAMPTZ,
 932        "TIMESTAMPLTZ": TokenType.TIMESTAMPLTZ,
 933        "TIMESTAMP_LTZ": TokenType.TIMESTAMPLTZ,
 934        "TIMESTAMPNTZ": TokenType.TIMESTAMPNTZ,
 935        "TIMESTAMP_NTZ": TokenType.TIMESTAMPNTZ,
 936        "DATE": TokenType.DATE,
 937        "DATETIME": TokenType.DATETIME,
 938        "INT4RANGE": TokenType.INT4RANGE,
 939        "INT4MULTIRANGE": TokenType.INT4MULTIRANGE,
 940        "INT8RANGE": TokenType.INT8RANGE,
 941        "INT8MULTIRANGE": TokenType.INT8MULTIRANGE,
 942        "NUMRANGE": TokenType.NUMRANGE,
 943        "NUMMULTIRANGE": TokenType.NUMMULTIRANGE,
 944        "TSRANGE": TokenType.TSRANGE,
 945        "TSMULTIRANGE": TokenType.TSMULTIRANGE,
 946        "TSTZRANGE": TokenType.TSTZRANGE,
 947        "TSTZMULTIRANGE": TokenType.TSTZMULTIRANGE,
 948        "DATERANGE": TokenType.DATERANGE,
 949        "DATEMULTIRANGE": TokenType.DATEMULTIRANGE,
 950        "UNIQUE": TokenType.UNIQUE,
 951        "VECTOR": TokenType.VECTOR,
 952        "STRUCT": TokenType.STRUCT,
 953        "SEQUENCE": TokenType.SEQUENCE,
 954        "VARIANT": TokenType.VARIANT,
 955        "ALTER": TokenType.ALTER,
 956        "ANALYZE": TokenType.ANALYZE,
 957        "CALL": TokenType.COMMAND,
 958        "COMMENT": TokenType.COMMENT,
 959        "EXPLAIN": TokenType.COMMAND,
 960        "GRANT": TokenType.GRANT,
 961        "OPTIMIZE": TokenType.COMMAND,
 962        "PREPARE": TokenType.COMMAND,
 963        "VACUUM": TokenType.COMMAND,
 964        "USER-DEFINED": TokenType.USERDEFINED,
 965        "FOR VERSION": TokenType.VERSION_SNAPSHOT,
 966        "FOR TIMESTAMP": TokenType.TIMESTAMP_SNAPSHOT,
 967    }
 968
 969    WHITE_SPACE: t.Dict[t.Optional[str], TokenType] = {
 970        " ": TokenType.SPACE,
 971        "\t": TokenType.SPACE,
 972        "\n": TokenType.BREAK,
 973        "\r": TokenType.BREAK,
 974    }
 975
 976    COMMANDS = {
 977        TokenType.COMMAND,
 978        TokenType.EXECUTE,
 979        TokenType.FETCH,
 980        TokenType.SHOW,
 981        TokenType.RENAME,
 982    }
 983
 984    COMMAND_PREFIX_TOKENS = {TokenType.SEMICOLON, TokenType.BEGIN}
 985
 986    # Handle numeric literals like in hive (3L = BIGINT)
 987    NUMERIC_LITERALS: t.Dict[str, str] = {}
 988
 989    COMMENTS = ["--", ("/*", "*/")]
 990
 991    __slots__ = (
 992        "sql",
 993        "size",
 994        "tokens",
 995        "dialect",
 996        "use_rs_tokenizer",
 997        "_start",
 998        "_current",
 999        "_line",
1000        "_col",
1001        "_comments",
1002        "_char",
1003        "_end",
1004        "_peek",
1005        "_prev_token_line",
1006        "_rs_dialect_settings",
1007    )
1008
1009    def __init__(
1010        self, dialect: DialectType = None, use_rs_tokenizer: t.Optional[bool] = None
1011    ) -> None:
1012        from sqlglot.dialects import Dialect
1013
1014        self.dialect = Dialect.get_or_raise(dialect)
1015
1016        # initialize `use_rs_tokenizer`, and allow it to be overwritten per Tokenizer instance
1017        self.use_rs_tokenizer = (
1018            use_rs_tokenizer if use_rs_tokenizer is not None else USE_RS_TOKENIZER
1019        )
1020
1021        if self.use_rs_tokenizer:
1022            self._rs_dialect_settings = RsTokenizerDialectSettings(
1023                unescaped_sequences=self.dialect.UNESCAPED_SEQUENCES,
1024                identifiers_can_start_with_digit=self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT,
1025                numbers_can_be_underscore_separated=self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED,
1026            )
1027
1028        self.reset()
1029
1030    def reset(self) -> None:
1031        self.sql = ""
1032        self.size = 0
1033        self.tokens: t.List[Token] = []
1034        self._start = 0
1035        self._current = 0
1036        self._line = 1
1037        self._col = 0
1038        self._comments: t.List[str] = []
1039
1040        self._char = ""
1041        self._end = False
1042        self._peek = ""
1043        self._prev_token_line = -1
1044
1045    def tokenize(self, sql: str) -> t.List[Token]:
1046        """Returns a list of tokens corresponding to the SQL string `sql`."""
1047        if self.use_rs_tokenizer:
1048            return self.tokenize_rs(sql)
1049
1050        self.reset()
1051        self.sql = sql
1052        self.size = len(sql)
1053
1054        try:
1055            self._scan()
1056        except Exception as e:
1057            start = max(self._current - 50, 0)
1058            end = min(self._current + 50, self.size - 1)
1059            context = self.sql[start:end]
1060            raise TokenError(f"Error tokenizing '{context}'") from e
1061
1062        return self.tokens
1063
1064    def _scan(self, until: t.Optional[t.Callable] = None) -> None:
1065        while self.size and not self._end:
1066            current = self._current
1067
1068            # Skip spaces here rather than iteratively calling advance() for performance reasons
1069            while current < self.size:
1070                char = self.sql[current]
1071
1072                if char.isspace() and (char == " " or char == "\t"):
1073                    current += 1
1074                else:
1075                    break
1076
1077            offset = current - self._current if current > self._current else 1
1078
1079            self._start = current
1080            self._advance(offset)
1081
1082            if not self._char.isspace():
1083                if self._char.isdigit():
1084                    self._scan_number()
1085                elif self._char in self._IDENTIFIERS:
1086                    self._scan_identifier(self._IDENTIFIERS[self._char])
1087                else:
1088                    self._scan_keywords()
1089
1090            if until and until():
1091                break
1092
1093        if self.tokens and self._comments:
1094            self.tokens[-1].comments.extend(self._comments)
1095
1096    def _chars(self, size: int) -> str:
1097        if size == 1:
1098            return self._char
1099
1100        start = self._current - 1
1101        end = start + size
1102
1103        return self.sql[start:end] if end <= self.size else ""
1104
1105    def _advance(self, i: int = 1, alnum: bool = False) -> None:
1106        if self.WHITE_SPACE.get(self._char) is TokenType.BREAK:
1107            # Ensures we don't count an extra line if we get a \r\n line break sequence
1108            if not (self._char == "\r" and self._peek == "\n"):
1109                self._col = 1
1110                self._line += 1
1111        else:
1112            self._col += i
1113
1114        self._current += i
1115        self._end = self._current >= self.size
1116        self._char = self.sql[self._current - 1]
1117        self._peek = "" if self._end else self.sql[self._current]
1118
1119        if alnum and self._char.isalnum():
1120            # Here we use local variables instead of attributes for better performance
1121            _col = self._col
1122            _current = self._current
1123            _end = self._end
1124            _peek = self._peek
1125
1126            while _peek.isalnum():
1127                _col += 1
1128                _current += 1
1129                _end = _current >= self.size
1130                _peek = "" if _end else self.sql[_current]
1131
1132            self._col = _col
1133            self._current = _current
1134            self._end = _end
1135            self._peek = _peek
1136            self._char = self.sql[_current - 1]
1137
1138    @property
1139    def _text(self) -> str:
1140        return self.sql[self._start : self._current]
1141
1142    def _add(self, token_type: TokenType, text: t.Optional[str] = None) -> None:
1143        self._prev_token_line = self._line
1144
1145        if self._comments and token_type == TokenType.SEMICOLON and self.tokens:
1146            self.tokens[-1].comments.extend(self._comments)
1147            self._comments = []
1148
1149        self.tokens.append(
1150            Token(
1151                token_type,
1152                text=self._text if text is None else text,
1153                line=self._line,
1154                col=self._col,
1155                start=self._start,
1156                end=self._current - 1,
1157                comments=self._comments,
1158            )
1159        )
1160        self._comments = []
1161
1162        # If we have either a semicolon or a begin token before the command's token, we'll parse
1163        # whatever follows the command's token as a string
1164        if (
1165            token_type in self.COMMANDS
1166            and self._peek != ";"
1167            and (len(self.tokens) == 1 or self.tokens[-2].token_type in self.COMMAND_PREFIX_TOKENS)
1168        ):
1169            start = self._current
1170            tokens = len(self.tokens)
1171            self._scan(lambda: self._peek == ";")
1172            self.tokens = self.tokens[:tokens]
1173            text = self.sql[start : self._current].strip()
1174            if text:
1175                self._add(TokenType.STRING, text)
1176
1177    def _scan_keywords(self) -> None:
1178        size = 0
1179        word = None
1180        chars = self._text
1181        char = chars
1182        prev_space = False
1183        skip = False
1184        trie = self._KEYWORD_TRIE
1185        single_token = char in self.SINGLE_TOKENS
1186
1187        while chars:
1188            if skip:
1189                result = TrieResult.PREFIX
1190            else:
1191                result, trie = in_trie(trie, char.upper())
1192
1193            if result == TrieResult.FAILED:
1194                break
1195            if result == TrieResult.EXISTS:
1196                word = chars
1197
1198            end = self._current + size
1199            size += 1
1200
1201            if end < self.size:
1202                char = self.sql[end]
1203                single_token = single_token or char in self.SINGLE_TOKENS
1204                is_space = char.isspace()
1205
1206                if not is_space or not prev_space:
1207                    if is_space:
1208                        char = " "
1209                    chars += char
1210                    prev_space = is_space
1211                    skip = False
1212                else:
1213                    skip = True
1214            else:
1215                char = ""
1216                break
1217
1218        if word:
1219            if self._scan_string(word):
1220                return
1221            if self._scan_comment(word):
1222                return
1223            if prev_space or single_token or not char:
1224                self._advance(size - 1)
1225                word = word.upper()
1226                self._add(self.KEYWORDS[word], text=word)
1227                return
1228
1229        if self._char in self.SINGLE_TOKENS:
1230            self._add(self.SINGLE_TOKENS[self._char], text=self._char)
1231            return
1232
1233        self._scan_var()
1234
1235    def _scan_comment(self, comment_start: str) -> bool:
1236        if comment_start not in self._COMMENTS:
1237            return False
1238
1239        comment_start_line = self._line
1240        comment_start_size = len(comment_start)
1241        comment_end = self._COMMENTS[comment_start]
1242
1243        if comment_end:
1244            # Skip the comment's start delimiter
1245            self._advance(comment_start_size)
1246
1247            comment_count = 1
1248            comment_end_size = len(comment_end)
1249
1250            while not self._end:
1251                if self._chars(comment_end_size) == comment_end:
1252                    comment_count -= 1
1253                    if not comment_count:
1254                        break
1255
1256                self._advance(alnum=True)
1257
1258                # Nested comments are allowed by some dialects, e.g. databricks, duckdb, postgres
1259                if (
1260                    self.NESTED_COMMENTS
1261                    and not self._end
1262                    and self._chars(comment_end_size) == comment_start
1263                ):
1264                    self._advance(comment_start_size)
1265                    comment_count += 1
1266
1267            self._comments.append(self._text[comment_start_size : -comment_end_size + 1])
1268            self._advance(comment_end_size - 1)
1269        else:
1270            while not self._end and self.WHITE_SPACE.get(self._peek) is not TokenType.BREAK:
1271                self._advance(alnum=True)
1272            self._comments.append(self._text[comment_start_size:])
1273
1274        if (
1275            comment_start == self.HINT_START
1276            and self.tokens
1277            and self.tokens[-1].token_type in self.TOKENS_PRECEDING_HINT
1278        ):
1279            self._add(TokenType.HINT)
1280
1281        # Leading comment is attached to the succeeding token, whilst trailing comment to the preceding.
1282        # Multiple consecutive comments are preserved by appending them to the current comments list.
1283        if comment_start_line == self._prev_token_line:
1284            self.tokens[-1].comments.extend(self._comments)
1285            self._comments = []
1286            self._prev_token_line = self._line
1287
1288        return True
1289
1290    def _scan_number(self) -> None:
1291        if self._char == "0":
1292            peek = self._peek.upper()
1293            if peek == "B":
1294                return self._scan_bits() if self.BIT_STRINGS else self._add(TokenType.NUMBER)
1295            elif peek == "X":
1296                return self._scan_hex() if self.HEX_STRINGS else self._add(TokenType.NUMBER)
1297
1298        decimal = False
1299        scientific = 0
1300
1301        while True:
1302            if self._peek.isdigit():
1303                self._advance()
1304            elif self._peek == "." and not decimal:
1305                if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER:
1306                    return self._add(TokenType.NUMBER)
1307                decimal = True
1308                self._advance()
1309            elif self._peek in ("-", "+") and scientific == 1:
1310                scientific += 1
1311                self._advance()
1312            elif self._peek.upper() == "E" and not scientific:
1313                scientific += 1
1314                self._advance()
1315            elif self._peek.isidentifier():
1316                number_text = self._text
1317                literal = ""
1318
1319                while self._peek.strip() and self._peek not in self.SINGLE_TOKENS:
1320                    literal += self._peek
1321                    self._advance()
1322
1323                token_type = self.KEYWORDS.get(self.NUMERIC_LITERALS.get(literal.upper(), ""))
1324
1325                if token_type:
1326                    self._add(TokenType.NUMBER, number_text)
1327                    self._add(TokenType.DCOLON, "::")
1328                    return self._add(token_type, literal)
1329                else:
1330                    replaced = literal.replace("_", "")
1331                    if self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED and replaced.isdigit():
1332                        return self._add(TokenType.NUMBER, number_text + replaced)
1333                    if self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT:
1334                        return self._add(TokenType.VAR)
1335
1336                self._advance(-len(literal))
1337                return self._add(TokenType.NUMBER, number_text)
1338            else:
1339                return self._add(TokenType.NUMBER)
1340
1341    def _scan_bits(self) -> None:
1342        self._advance()
1343        value = self._extract_value()
1344        try:
1345            # If `value` can't be converted to a binary, fallback to tokenizing it as an identifier
1346            int(value, 2)
1347            self._add(TokenType.BIT_STRING, value[2:])  # Drop the 0b
1348        except ValueError:
1349            self._add(TokenType.IDENTIFIER)
1350
1351    def _scan_hex(self) -> None:
1352        self._advance()
1353        value = self._extract_value()
1354        try:
1355            # If `value` can't be converted to a hex, fallback to tokenizing it as an identifier
1356            int(value, 16)
1357            self._add(TokenType.HEX_STRING, value[2:])  # Drop the 0x
1358        except ValueError:
1359            self._add(TokenType.IDENTIFIER)
1360
1361    def _extract_value(self) -> str:
1362        while True:
1363            char = self._peek.strip()
1364            if char and char not in self.SINGLE_TOKENS:
1365                self._advance(alnum=True)
1366            else:
1367                break
1368
1369        return self._text
1370
1371    def _scan_string(self, start: str) -> bool:
1372        base = None
1373        token_type = TokenType.STRING
1374
1375        if start in self._QUOTES:
1376            end = self._QUOTES[start]
1377        elif start in self._FORMAT_STRINGS:
1378            end, token_type = self._FORMAT_STRINGS[start]
1379
1380            if token_type == TokenType.HEX_STRING:
1381                base = 16
1382            elif token_type == TokenType.BIT_STRING:
1383                base = 2
1384            elif token_type == TokenType.HEREDOC_STRING:
1385                self._advance()
1386
1387                if self._char == end:
1388                    tag = ""
1389                else:
1390                    tag = self._extract_string(
1391                        end,
1392                        raw_string=True,
1393                        raise_unmatched=not self.HEREDOC_TAG_IS_IDENTIFIER,
1394                    )
1395
1396                if tag and self.HEREDOC_TAG_IS_IDENTIFIER and (self._end or not tag.isidentifier()):
1397                    if not self._end:
1398                        self._advance(-1)
1399
1400                    self._advance(-len(tag))
1401                    self._add(self.HEREDOC_STRING_ALTERNATIVE)
1402                    return True
1403
1404                end = f"{start}{tag}{end}"
1405        else:
1406            return False
1407
1408        self._advance(len(start))
1409        text = self._extract_string(end, raw_string=token_type == TokenType.RAW_STRING)
1410
1411        if base:
1412            try:
1413                int(text, base)
1414            except Exception:
1415                raise TokenError(
1416                    f"Numeric string contains invalid characters from {self._line}:{self._start}"
1417                )
1418
1419        self._add(token_type, text)
1420        return True
1421
1422    def _scan_identifier(self, identifier_end: str) -> None:
1423        self._advance()
1424        text = self._extract_string(
1425            identifier_end, escapes=self._IDENTIFIER_ESCAPES | {identifier_end}
1426        )
1427        self._add(TokenType.IDENTIFIER, text)
1428
1429    def _scan_var(self) -> None:
1430        while True:
1431            char = self._peek.strip()
1432            if char and (char in self.VAR_SINGLE_TOKENS or char not in self.SINGLE_TOKENS):
1433                self._advance(alnum=True)
1434            else:
1435                break
1436
1437        self._add(
1438            TokenType.VAR
1439            if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER
1440            else self.KEYWORDS.get(self._text.upper(), TokenType.VAR)
1441        )
1442
1443    def _extract_string(
1444        self,
1445        delimiter: str,
1446        escapes: t.Optional[t.Set[str]] = None,
1447        raw_string: bool = False,
1448        raise_unmatched: bool = True,
1449    ) -> str:
1450        text = ""
1451        delim_size = len(delimiter)
1452        escapes = self._STRING_ESCAPES if escapes is None else escapes
1453
1454        while True:
1455            if (
1456                not raw_string
1457                and self.dialect.UNESCAPED_SEQUENCES
1458                and self._peek
1459                and self._char in self.STRING_ESCAPES
1460            ):
1461                unescaped_sequence = self.dialect.UNESCAPED_SEQUENCES.get(self._char + self._peek)
1462                if unescaped_sequence:
1463                    self._advance(2)
1464                    text += unescaped_sequence
1465                    continue
1466            if (
1467                (self.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS or not raw_string)
1468                and self._char in escapes
1469                and (self._peek == delimiter or self._peek in escapes)
1470                and (self._char not in self._QUOTES or self._char == self._peek)
1471            ):
1472                if self._peek == delimiter:
1473                    text += self._peek
1474                else:
1475                    text += self._char + self._peek
1476
1477                if self._current + 1 < self.size:
1478                    self._advance(2)
1479                else:
1480                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._current}")
1481            else:
1482                if self._chars(delim_size) == delimiter:
1483                    if delim_size > 1:
1484                        self._advance(delim_size - 1)
1485                    break
1486
1487                if self._end:
1488                    if not raise_unmatched:
1489                        return text + self._char
1490
1491                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._start}")
1492
1493                current = self._current - 1
1494                self._advance(alnum=True)
1495                text += self.sql[current : self._current - 1]
1496
1497        return text
1498
1499    def tokenize_rs(self, sql: str) -> t.List[Token]:
1500        if not self._RS_TOKENIZER:
1501            raise SqlglotError("Rust tokenizer is not available")
1502
1503        try:
1504            tokens = self._RS_TOKENIZER.tokenize(sql, self._rs_dialect_settings)
1505            for token in tokens:
1506                token.token_type = _ALL_TOKEN_TYPES[token.token_type_index]
1507            return tokens
1508        except Exception as e:
1509            raise TokenError(str(e))