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    VOID = auto()
 226
 227    # keywords
 228    ALIAS = auto()
 229    ALTER = auto()
 230    ALWAYS = auto()
 231    ALL = auto()
 232    ANTI = auto()
 233    ANY = auto()
 234    APPLY = auto()
 235    ARRAY = auto()
 236    ASC = auto()
 237    ASOF = auto()
 238    ATTACH = auto()
 239    AUTO_INCREMENT = auto()
 240    BEGIN = auto()
 241    BETWEEN = auto()
 242    BULK_COLLECT_INTO = auto()
 243    CACHE = auto()
 244    CASE = auto()
 245    CHARACTER_SET = auto()
 246    CLUSTER_BY = auto()
 247    COLLATE = auto()
 248    COMMAND = auto()
 249    COMMENT = auto()
 250    COMMIT = auto()
 251    CONNECT_BY = auto()
 252    CONSTRAINT = auto()
 253    COPY = auto()
 254    CREATE = auto()
 255    CROSS = auto()
 256    CUBE = auto()
 257    CURRENT_DATE = auto()
 258    CURRENT_DATETIME = auto()
 259    CURRENT_SCHEMA = auto()
 260    CURRENT_TIME = auto()
 261    CURRENT_TIMESTAMP = auto()
 262    CURRENT_USER = auto()
 263    DECLARE = auto()
 264    DEFAULT = auto()
 265    DELETE = auto()
 266    DESC = auto()
 267    DESCRIBE = auto()
 268    DETACH = auto()
 269    DICTIONARY = auto()
 270    DISTINCT = auto()
 271    DISTRIBUTE_BY = auto()
 272    DIV = auto()
 273    DROP = auto()
 274    ELSE = auto()
 275    END = auto()
 276    ESCAPE = auto()
 277    EXCEPT = auto()
 278    EXECUTE = auto()
 279    EXISTS = auto()
 280    FALSE = auto()
 281    FETCH = auto()
 282    FILE_FORMAT = auto()
 283    FILTER = auto()
 284    FINAL = auto()
 285    FIRST = auto()
 286    FOR = auto()
 287    FORCE = auto()
 288    FOREIGN_KEY = auto()
 289    FORMAT = auto()
 290    FROM = auto()
 291    FULL = auto()
 292    FUNCTION = auto()
 293    GET = auto()
 294    GLOB = auto()
 295    GLOBAL = auto()
 296    GRANT = auto()
 297    GROUP_BY = auto()
 298    GROUPING_SETS = auto()
 299    HAVING = auto()
 300    HINT = auto()
 301    IGNORE = auto()
 302    ILIKE = auto()
 303    ILIKE_ANY = auto()
 304    IN = auto()
 305    INDEX = auto()
 306    INNER = auto()
 307    INSERT = auto()
 308    INTERSECT = auto()
 309    INTERVAL = auto()
 310    INTO = auto()
 311    INTRODUCER = auto()
 312    IRLIKE = auto()
 313    IS = auto()
 314    ISNULL = auto()
 315    JOIN = auto()
 316    JOIN_MARKER = auto()
 317    KEEP = auto()
 318    KEY = auto()
 319    KILL = auto()
 320    LANGUAGE = auto()
 321    LATERAL = auto()
 322    LEFT = auto()
 323    LIKE = auto()
 324    LIKE_ANY = auto()
 325    LIMIT = auto()
 326    LIST = auto()
 327    LOAD = auto()
 328    LOCK = auto()
 329    MAP = auto()
 330    MATCH_CONDITION = auto()
 331    MATCH_RECOGNIZE = auto()
 332    MEMBER_OF = auto()
 333    MERGE = auto()
 334    MOD = auto()
 335    MODEL = auto()
 336    NATURAL = auto()
 337    NEXT = auto()
 338    NOTHING = auto()
 339    NOTNULL = auto()
 340    NULL = auto()
 341    OBJECT_IDENTIFIER = auto()
 342    OFFSET = auto()
 343    ON = auto()
 344    ONLY = auto()
 345    OPERATOR = auto()
 346    ORDER_BY = auto()
 347    ORDER_SIBLINGS_BY = auto()
 348    ORDERED = auto()
 349    ORDINALITY = auto()
 350    OUTER = auto()
 351    OVER = auto()
 352    OVERLAPS = auto()
 353    OVERWRITE = auto()
 354    PARTITION = auto()
 355    PARTITION_BY = auto()
 356    PERCENT = auto()
 357    PIVOT = auto()
 358    PLACEHOLDER = auto()
 359    POSITIONAL = auto()
 360    PRAGMA = auto()
 361    PREWHERE = auto()
 362    PRIMARY_KEY = auto()
 363    PROCEDURE = auto()
 364    PROPERTIES = auto()
 365    PSEUDO_TYPE = auto()
 366    PUT = auto()
 367    QUALIFY = auto()
 368    QUOTE = auto()
 369    RANGE = auto()
 370    RECURSIVE = auto()
 371    REFRESH = auto()
 372    RENAME = auto()
 373    REPLACE = auto()
 374    RETURNING = auto()
 375    REFERENCES = auto()
 376    RIGHT = auto()
 377    RLIKE = auto()
 378    ROLLBACK = auto()
 379    ROLLUP = auto()
 380    ROW = auto()
 381    ROWS = auto()
 382    SELECT = auto()
 383    SEMI = auto()
 384    SEPARATOR = auto()
 385    SEQUENCE = auto()
 386    SERDE_PROPERTIES = auto()
 387    SET = auto()
 388    SETTINGS = auto()
 389    SHOW = auto()
 390    SIMILAR_TO = auto()
 391    SOME = auto()
 392    SORT_BY = auto()
 393    START_WITH = auto()
 394    STORAGE_INTEGRATION = auto()
 395    STRAIGHT_JOIN = auto()
 396    STRUCT = auto()
 397    SUMMARIZE = auto()
 398    TABLE_SAMPLE = auto()
 399    TAG = auto()
 400    TEMPORARY = auto()
 401    TOP = auto()
 402    THEN = auto()
 403    TRUE = auto()
 404    TRUNCATE = auto()
 405    UNCACHE = auto()
 406    UNION = auto()
 407    UNNEST = auto()
 408    UNPIVOT = auto()
 409    UPDATE = auto()
 410    USE = auto()
 411    USING = auto()
 412    VALUES = auto()
 413    VIEW = auto()
 414    VOLATILE = auto()
 415    WHEN = auto()
 416    WHERE = auto()
 417    WINDOW = auto()
 418    WITH = auto()
 419    UNIQUE = auto()
 420    VERSION_SNAPSHOT = auto()
 421    TIMESTAMP_SNAPSHOT = auto()
 422    OPTION = auto()
 423    SINK = auto()
 424    SOURCE = auto()
 425    ANALYZE = auto()
 426    NAMESPACE = auto()
 427    EXPORT = auto()
 428
 429
 430_ALL_TOKEN_TYPES = list(TokenType)
 431_TOKEN_TYPE_TO_INDEX = {token_type: i for i, token_type in enumerate(_ALL_TOKEN_TYPES)}
 432
 433
 434class Token:
 435    __slots__ = ("token_type", "text", "line", "col", "start", "end", "comments")
 436
 437    @classmethod
 438    def number(cls, number: int) -> Token:
 439        """Returns a NUMBER token with `number` as its text."""
 440        return cls(TokenType.NUMBER, str(number))
 441
 442    @classmethod
 443    def string(cls, string: str) -> Token:
 444        """Returns a STRING token with `string` as its text."""
 445        return cls(TokenType.STRING, string)
 446
 447    @classmethod
 448    def identifier(cls, identifier: str) -> Token:
 449        """Returns an IDENTIFIER token with `identifier` as its text."""
 450        return cls(TokenType.IDENTIFIER, identifier)
 451
 452    @classmethod
 453    def var(cls, var: str) -> Token:
 454        """Returns an VAR token with `var` as its text."""
 455        return cls(TokenType.VAR, var)
 456
 457    def __init__(
 458        self,
 459        token_type: TokenType,
 460        text: str,
 461        line: int = 1,
 462        col: int = 1,
 463        start: int = 0,
 464        end: int = 0,
 465        comments: t.Optional[t.List[str]] = None,
 466    ) -> None:
 467        """Token initializer.
 468
 469        Args:
 470            token_type: The TokenType Enum.
 471            text: The text of the token.
 472            line: The line that the token ends on.
 473            col: The column that the token ends on.
 474            start: The start index of the token.
 475            end: The ending index of the token.
 476            comments: The comments to attach to the token.
 477        """
 478        self.token_type = token_type
 479        self.text = text
 480        self.line = line
 481        self.col = col
 482        self.start = start
 483        self.end = end
 484        self.comments = [] if comments is None else comments
 485
 486    def __repr__(self) -> str:
 487        attributes = ", ".join(f"{k}: {getattr(self, k)}" for k in self.__slots__)
 488        return f"<Token {attributes}>"
 489
 490
 491class _Tokenizer(type):
 492    def __new__(cls, clsname, bases, attrs):
 493        klass = super().__new__(cls, clsname, bases, attrs)
 494
 495        def _convert_quotes(arr: t.List[str | t.Tuple[str, str]]) -> t.Dict[str, str]:
 496            return dict(
 497                (item, item) if isinstance(item, str) else (item[0], item[1]) for item in arr
 498            )
 499
 500        def _quotes_to_format(
 501            token_type: TokenType, arr: t.List[str | t.Tuple[str, str]]
 502        ) -> t.Dict[str, t.Tuple[str, TokenType]]:
 503            return {k: (v, token_type) for k, v in _convert_quotes(arr).items()}
 504
 505        klass._QUOTES = _convert_quotes(klass.QUOTES)
 506        klass._IDENTIFIERS = _convert_quotes(klass.IDENTIFIERS)
 507
 508        klass._FORMAT_STRINGS = {
 509            **{
 510                p + s: (e, TokenType.NATIONAL_STRING)
 511                for s, e in klass._QUOTES.items()
 512                for p in ("n", "N")
 513            },
 514            **_quotes_to_format(TokenType.BIT_STRING, klass.BIT_STRINGS),
 515            **_quotes_to_format(TokenType.BYTE_STRING, klass.BYTE_STRINGS),
 516            **_quotes_to_format(TokenType.HEX_STRING, klass.HEX_STRINGS),
 517            **_quotes_to_format(TokenType.RAW_STRING, klass.RAW_STRINGS),
 518            **_quotes_to_format(TokenType.HEREDOC_STRING, klass.HEREDOC_STRINGS),
 519            **_quotes_to_format(TokenType.UNICODE_STRING, klass.UNICODE_STRINGS),
 520        }
 521
 522        klass._STRING_ESCAPES = set(klass.STRING_ESCAPES)
 523        klass._IDENTIFIER_ESCAPES = set(klass.IDENTIFIER_ESCAPES)
 524        klass._COMMENTS = {
 525            **dict(
 526                (comment, None) if isinstance(comment, str) else (comment[0], comment[1])
 527                for comment in klass.COMMENTS
 528            ),
 529            "{#": "#}",  # Ensure Jinja comments are tokenized correctly in all dialects
 530        }
 531        if klass.HINT_START in klass.KEYWORDS:
 532            klass._COMMENTS[klass.HINT_START] = "*/"
 533
 534        klass._KEYWORD_TRIE = new_trie(
 535            key.upper()
 536            for key in (
 537                *klass.KEYWORDS,
 538                *klass._COMMENTS,
 539                *klass._QUOTES,
 540                *klass._FORMAT_STRINGS,
 541            )
 542            if " " in key or any(single in key for single in klass.SINGLE_TOKENS)
 543        )
 544
 545        if USE_RS_TOKENIZER:
 546            settings = RsTokenizerSettings(
 547                white_space={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.WHITE_SPACE.items()},
 548                single_tokens={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.SINGLE_TOKENS.items()},
 549                keywords={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.KEYWORDS.items()},
 550                numeric_literals=klass.NUMERIC_LITERALS,
 551                identifiers=klass._IDENTIFIERS,
 552                identifier_escapes=klass._IDENTIFIER_ESCAPES,
 553                string_escapes=klass._STRING_ESCAPES,
 554                quotes=klass._QUOTES,
 555                format_strings={
 556                    k: (v1, _TOKEN_TYPE_TO_INDEX[v2])
 557                    for k, (v1, v2) in klass._FORMAT_STRINGS.items()
 558                },
 559                has_bit_strings=bool(klass.BIT_STRINGS),
 560                has_hex_strings=bool(klass.HEX_STRINGS),
 561                comments=klass._COMMENTS,
 562                var_single_tokens=klass.VAR_SINGLE_TOKENS,
 563                commands={_TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMANDS},
 564                command_prefix_tokens={
 565                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMAND_PREFIX_TOKENS
 566                },
 567                heredoc_tag_is_identifier=klass.HEREDOC_TAG_IS_IDENTIFIER,
 568                string_escapes_allowed_in_raw_strings=klass.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS,
 569                nested_comments=klass.NESTED_COMMENTS,
 570                hint_start=klass.HINT_START,
 571                tokens_preceding_hint={
 572                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.TOKENS_PRECEDING_HINT
 573                },
 574            )
 575            token_types = RsTokenTypeSettings(
 576                bit_string=_TOKEN_TYPE_TO_INDEX[TokenType.BIT_STRING],
 577                break_=_TOKEN_TYPE_TO_INDEX[TokenType.BREAK],
 578                dcolon=_TOKEN_TYPE_TO_INDEX[TokenType.DCOLON],
 579                heredoc_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEREDOC_STRING],
 580                raw_string=_TOKEN_TYPE_TO_INDEX[TokenType.RAW_STRING],
 581                hex_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEX_STRING],
 582                identifier=_TOKEN_TYPE_TO_INDEX[TokenType.IDENTIFIER],
 583                number=_TOKEN_TYPE_TO_INDEX[TokenType.NUMBER],
 584                parameter=_TOKEN_TYPE_TO_INDEX[TokenType.PARAMETER],
 585                semicolon=_TOKEN_TYPE_TO_INDEX[TokenType.SEMICOLON],
 586                string=_TOKEN_TYPE_TO_INDEX[TokenType.STRING],
 587                var=_TOKEN_TYPE_TO_INDEX[TokenType.VAR],
 588                heredoc_string_alternative=_TOKEN_TYPE_TO_INDEX[klass.HEREDOC_STRING_ALTERNATIVE],
 589                hint=_TOKEN_TYPE_TO_INDEX[TokenType.HINT],
 590            )
 591            klass._RS_TOKENIZER = RsTokenizer(settings, token_types)
 592        else:
 593            klass._RS_TOKENIZER = None
 594
 595        return klass
 596
 597
 598class Tokenizer(metaclass=_Tokenizer):
 599    SINGLE_TOKENS = {
 600        "(": TokenType.L_PAREN,
 601        ")": TokenType.R_PAREN,
 602        "[": TokenType.L_BRACKET,
 603        "]": TokenType.R_BRACKET,
 604        "{": TokenType.L_BRACE,
 605        "}": TokenType.R_BRACE,
 606        "&": TokenType.AMP,
 607        "^": TokenType.CARET,
 608        ":": TokenType.COLON,
 609        ",": TokenType.COMMA,
 610        ".": TokenType.DOT,
 611        "-": TokenType.DASH,
 612        "=": TokenType.EQ,
 613        ">": TokenType.GT,
 614        "<": TokenType.LT,
 615        "%": TokenType.MOD,
 616        "!": TokenType.NOT,
 617        "|": TokenType.PIPE,
 618        "+": TokenType.PLUS,
 619        ";": TokenType.SEMICOLON,
 620        "/": TokenType.SLASH,
 621        "\\": TokenType.BACKSLASH,
 622        "*": TokenType.STAR,
 623        "~": TokenType.TILDA,
 624        "?": TokenType.PLACEHOLDER,
 625        "@": TokenType.PARAMETER,
 626        "#": TokenType.HASH,
 627        # Used for breaking a var like x'y' but nothing else the token type doesn't matter
 628        "'": TokenType.UNKNOWN,
 629        "`": TokenType.UNKNOWN,
 630        '"': TokenType.UNKNOWN,
 631    }
 632
 633    BIT_STRINGS: t.List[str | t.Tuple[str, str]] = []
 634    BYTE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 635    HEX_STRINGS: t.List[str | t.Tuple[str, str]] = []
 636    RAW_STRINGS: t.List[str | t.Tuple[str, str]] = []
 637    HEREDOC_STRINGS: t.List[str | t.Tuple[str, str]] = []
 638    UNICODE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 639    IDENTIFIERS: t.List[str | t.Tuple[str, str]] = ['"']
 640    QUOTES: t.List[t.Tuple[str, str] | str] = ["'"]
 641    STRING_ESCAPES = ["'"]
 642    VAR_SINGLE_TOKENS: t.Set[str] = set()
 643
 644    # The strings in this list can always be used as escapes, regardless of the surrounding
 645    # identifier delimiters. By default, the closing delimiter is assumed to also act as an
 646    # identifier escape, e.g. if we use double-quotes, then they also act as escapes: "x"""
 647    IDENTIFIER_ESCAPES: t.List[str] = []
 648
 649    # Whether the heredoc tags follow the same lexical rules as unquoted identifiers
 650    HEREDOC_TAG_IS_IDENTIFIER = False
 651
 652    # Token that we'll generate as a fallback if the heredoc prefix doesn't correspond to a heredoc
 653    HEREDOC_STRING_ALTERNATIVE = TokenType.VAR
 654
 655    # Whether string escape characters function as such when placed within raw strings
 656    STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS = True
 657
 658    NESTED_COMMENTS = True
 659
 660    HINT_START = "/*+"
 661
 662    TOKENS_PRECEDING_HINT = {TokenType.SELECT, TokenType.INSERT, TokenType.UPDATE, TokenType.DELETE}
 663
 664    # Autofilled
 665    _COMMENTS: t.Dict[str, str] = {}
 666    _FORMAT_STRINGS: t.Dict[str, t.Tuple[str, TokenType]] = {}
 667    _IDENTIFIERS: t.Dict[str, str] = {}
 668    _IDENTIFIER_ESCAPES: t.Set[str] = set()
 669    _QUOTES: t.Dict[str, str] = {}
 670    _STRING_ESCAPES: t.Set[str] = set()
 671    _KEYWORD_TRIE: t.Dict = {}
 672    _RS_TOKENIZER: t.Optional[t.Any] = None
 673
 674    KEYWORDS: t.Dict[str, TokenType] = {
 675        **{f"{{%{postfix}": TokenType.BLOCK_START for postfix in ("", "+", "-")},
 676        **{f"{prefix}%}}": TokenType.BLOCK_END for prefix in ("", "+", "-")},
 677        **{f"{{{{{postfix}": TokenType.BLOCK_START for postfix in ("+", "-")},
 678        **{f"{prefix}}}}}": TokenType.BLOCK_END for prefix in ("+", "-")},
 679        HINT_START: TokenType.HINT,
 680        "==": TokenType.EQ,
 681        "::": TokenType.DCOLON,
 682        "||": TokenType.DPIPE,
 683        ">=": TokenType.GTE,
 684        "<=": TokenType.LTE,
 685        "<>": TokenType.NEQ,
 686        "!=": TokenType.NEQ,
 687        ":=": TokenType.COLON_EQ,
 688        "<=>": TokenType.NULLSAFE_EQ,
 689        "->": TokenType.ARROW,
 690        "->>": TokenType.DARROW,
 691        "=>": TokenType.FARROW,
 692        "#>": TokenType.HASH_ARROW,
 693        "#>>": TokenType.DHASH_ARROW,
 694        "<->": TokenType.LR_ARROW,
 695        "&&": TokenType.DAMP,
 696        "??": TokenType.DQMARK,
 697        "~~~": TokenType.GLOB,
 698        "~~": TokenType.LIKE,
 699        "~~*": TokenType.ILIKE,
 700        "~*": TokenType.IRLIKE,
 701        "ALL": TokenType.ALL,
 702        "ALWAYS": TokenType.ALWAYS,
 703        "AND": TokenType.AND,
 704        "ANTI": TokenType.ANTI,
 705        "ANY": TokenType.ANY,
 706        "ASC": TokenType.ASC,
 707        "AS": TokenType.ALIAS,
 708        "ASOF": TokenType.ASOF,
 709        "AUTOINCREMENT": TokenType.AUTO_INCREMENT,
 710        "AUTO_INCREMENT": TokenType.AUTO_INCREMENT,
 711        "BEGIN": TokenType.BEGIN,
 712        "BETWEEN": TokenType.BETWEEN,
 713        "CACHE": TokenType.CACHE,
 714        "UNCACHE": TokenType.UNCACHE,
 715        "CASE": TokenType.CASE,
 716        "CHARACTER SET": TokenType.CHARACTER_SET,
 717        "CLUSTER BY": TokenType.CLUSTER_BY,
 718        "COLLATE": TokenType.COLLATE,
 719        "COLUMN": TokenType.COLUMN,
 720        "COMMIT": TokenType.COMMIT,
 721        "CONNECT BY": TokenType.CONNECT_BY,
 722        "CONSTRAINT": TokenType.CONSTRAINT,
 723        "COPY": TokenType.COPY,
 724        "CREATE": TokenType.CREATE,
 725        "CROSS": TokenType.CROSS,
 726        "CUBE": TokenType.CUBE,
 727        "CURRENT_DATE": TokenType.CURRENT_DATE,
 728        "CURRENT_SCHEMA": TokenType.CURRENT_SCHEMA,
 729        "CURRENT_TIME": TokenType.CURRENT_TIME,
 730        "CURRENT_TIMESTAMP": TokenType.CURRENT_TIMESTAMP,
 731        "CURRENT_USER": TokenType.CURRENT_USER,
 732        "DATABASE": TokenType.DATABASE,
 733        "DEFAULT": TokenType.DEFAULT,
 734        "DELETE": TokenType.DELETE,
 735        "DESC": TokenType.DESC,
 736        "DESCRIBE": TokenType.DESCRIBE,
 737        "DISTINCT": TokenType.DISTINCT,
 738        "DISTRIBUTE BY": TokenType.DISTRIBUTE_BY,
 739        "DIV": TokenType.DIV,
 740        "DROP": TokenType.DROP,
 741        "ELSE": TokenType.ELSE,
 742        "END": TokenType.END,
 743        "ENUM": TokenType.ENUM,
 744        "ESCAPE": TokenType.ESCAPE,
 745        "EXCEPT": TokenType.EXCEPT,
 746        "EXECUTE": TokenType.EXECUTE,
 747        "EXISTS": TokenType.EXISTS,
 748        "FALSE": TokenType.FALSE,
 749        "FETCH": TokenType.FETCH,
 750        "FILTER": TokenType.FILTER,
 751        "FIRST": TokenType.FIRST,
 752        "FULL": TokenType.FULL,
 753        "FUNCTION": TokenType.FUNCTION,
 754        "FOR": TokenType.FOR,
 755        "FOREIGN KEY": TokenType.FOREIGN_KEY,
 756        "FORMAT": TokenType.FORMAT,
 757        "FROM": TokenType.FROM,
 758        "GEOGRAPHY": TokenType.GEOGRAPHY,
 759        "GEOMETRY": TokenType.GEOMETRY,
 760        "GLOB": TokenType.GLOB,
 761        "GROUP BY": TokenType.GROUP_BY,
 762        "GROUPING SETS": TokenType.GROUPING_SETS,
 763        "HAVING": TokenType.HAVING,
 764        "ILIKE": TokenType.ILIKE,
 765        "IN": TokenType.IN,
 766        "INDEX": TokenType.INDEX,
 767        "INET": TokenType.INET,
 768        "INNER": TokenType.INNER,
 769        "INSERT": TokenType.INSERT,
 770        "INTERVAL": TokenType.INTERVAL,
 771        "INTERSECT": TokenType.INTERSECT,
 772        "INTO": TokenType.INTO,
 773        "IS": TokenType.IS,
 774        "ISNULL": TokenType.ISNULL,
 775        "JOIN": TokenType.JOIN,
 776        "KEEP": TokenType.KEEP,
 777        "KILL": TokenType.KILL,
 778        "LATERAL": TokenType.LATERAL,
 779        "LEFT": TokenType.LEFT,
 780        "LIKE": TokenType.LIKE,
 781        "LIMIT": TokenType.LIMIT,
 782        "LOAD": TokenType.LOAD,
 783        "LOCK": TokenType.LOCK,
 784        "MERGE": TokenType.MERGE,
 785        "NAMESPACE": TokenType.NAMESPACE,
 786        "NATURAL": TokenType.NATURAL,
 787        "NEXT": TokenType.NEXT,
 788        "NOT": TokenType.NOT,
 789        "NOTNULL": TokenType.NOTNULL,
 790        "NULL": TokenType.NULL,
 791        "OBJECT": TokenType.OBJECT,
 792        "OFFSET": TokenType.OFFSET,
 793        "ON": TokenType.ON,
 794        "OR": TokenType.OR,
 795        "XOR": TokenType.XOR,
 796        "ORDER BY": TokenType.ORDER_BY,
 797        "ORDINALITY": TokenType.ORDINALITY,
 798        "OUTER": TokenType.OUTER,
 799        "OVER": TokenType.OVER,
 800        "OVERLAPS": TokenType.OVERLAPS,
 801        "OVERWRITE": TokenType.OVERWRITE,
 802        "PARTITION": TokenType.PARTITION,
 803        "PARTITION BY": TokenType.PARTITION_BY,
 804        "PARTITIONED BY": TokenType.PARTITION_BY,
 805        "PARTITIONED_BY": TokenType.PARTITION_BY,
 806        "PERCENT": TokenType.PERCENT,
 807        "PIVOT": TokenType.PIVOT,
 808        "PRAGMA": TokenType.PRAGMA,
 809        "PRIMARY KEY": TokenType.PRIMARY_KEY,
 810        "PROCEDURE": TokenType.PROCEDURE,
 811        "QUALIFY": TokenType.QUALIFY,
 812        "RANGE": TokenType.RANGE,
 813        "RECURSIVE": TokenType.RECURSIVE,
 814        "REGEXP": TokenType.RLIKE,
 815        "RENAME": TokenType.RENAME,
 816        "REPLACE": TokenType.REPLACE,
 817        "RETURNING": TokenType.RETURNING,
 818        "REFERENCES": TokenType.REFERENCES,
 819        "RIGHT": TokenType.RIGHT,
 820        "RLIKE": TokenType.RLIKE,
 821        "ROLLBACK": TokenType.ROLLBACK,
 822        "ROLLUP": TokenType.ROLLUP,
 823        "ROW": TokenType.ROW,
 824        "ROWS": TokenType.ROWS,
 825        "SCHEMA": TokenType.SCHEMA,
 826        "SELECT": TokenType.SELECT,
 827        "SEMI": TokenType.SEMI,
 828        "SET": TokenType.SET,
 829        "SETTINGS": TokenType.SETTINGS,
 830        "SHOW": TokenType.SHOW,
 831        "SIMILAR TO": TokenType.SIMILAR_TO,
 832        "SOME": TokenType.SOME,
 833        "SORT BY": TokenType.SORT_BY,
 834        "START WITH": TokenType.START_WITH,
 835        "STRAIGHT_JOIN": TokenType.STRAIGHT_JOIN,
 836        "TABLE": TokenType.TABLE,
 837        "TABLESAMPLE": TokenType.TABLE_SAMPLE,
 838        "TEMP": TokenType.TEMPORARY,
 839        "TEMPORARY": TokenType.TEMPORARY,
 840        "THEN": TokenType.THEN,
 841        "TRUE": TokenType.TRUE,
 842        "TRUNCATE": TokenType.TRUNCATE,
 843        "UNION": TokenType.UNION,
 844        "UNKNOWN": TokenType.UNKNOWN,
 845        "UNNEST": TokenType.UNNEST,
 846        "UNPIVOT": TokenType.UNPIVOT,
 847        "UPDATE": TokenType.UPDATE,
 848        "USE": TokenType.USE,
 849        "USING": TokenType.USING,
 850        "UUID": TokenType.UUID,
 851        "VALUES": TokenType.VALUES,
 852        "VIEW": TokenType.VIEW,
 853        "VOLATILE": TokenType.VOLATILE,
 854        "WHEN": TokenType.WHEN,
 855        "WHERE": TokenType.WHERE,
 856        "WINDOW": TokenType.WINDOW,
 857        "WITH": TokenType.WITH,
 858        "APPLY": TokenType.APPLY,
 859        "ARRAY": TokenType.ARRAY,
 860        "BIT": TokenType.BIT,
 861        "BOOL": TokenType.BOOLEAN,
 862        "BOOLEAN": TokenType.BOOLEAN,
 863        "BYTE": TokenType.TINYINT,
 864        "MEDIUMINT": TokenType.MEDIUMINT,
 865        "INT1": TokenType.TINYINT,
 866        "TINYINT": TokenType.TINYINT,
 867        "INT16": TokenType.SMALLINT,
 868        "SHORT": TokenType.SMALLINT,
 869        "SMALLINT": TokenType.SMALLINT,
 870        "HUGEINT": TokenType.INT128,
 871        "UHUGEINT": TokenType.UINT128,
 872        "INT2": TokenType.SMALLINT,
 873        "INTEGER": TokenType.INT,
 874        "INT": TokenType.INT,
 875        "INT4": TokenType.INT,
 876        "INT32": TokenType.INT,
 877        "INT64": TokenType.BIGINT,
 878        "INT128": TokenType.INT128,
 879        "INT256": TokenType.INT256,
 880        "LONG": TokenType.BIGINT,
 881        "BIGINT": TokenType.BIGINT,
 882        "INT8": TokenType.TINYINT,
 883        "UINT": TokenType.UINT,
 884        "UINT128": TokenType.UINT128,
 885        "UINT256": TokenType.UINT256,
 886        "DEC": TokenType.DECIMAL,
 887        "DECIMAL": TokenType.DECIMAL,
 888        "DECIMAL32": TokenType.DECIMAL32,
 889        "DECIMAL64": TokenType.DECIMAL64,
 890        "DECIMAL128": TokenType.DECIMAL128,
 891        "DECIMAL256": TokenType.DECIMAL256,
 892        "BIGDECIMAL": TokenType.BIGDECIMAL,
 893        "BIGNUMERIC": TokenType.BIGDECIMAL,
 894        "LIST": TokenType.LIST,
 895        "MAP": TokenType.MAP,
 896        "NULLABLE": TokenType.NULLABLE,
 897        "NUMBER": TokenType.DECIMAL,
 898        "NUMERIC": TokenType.DECIMAL,
 899        "FIXED": TokenType.DECIMAL,
 900        "REAL": TokenType.FLOAT,
 901        "FLOAT": TokenType.FLOAT,
 902        "FLOAT4": TokenType.FLOAT,
 903        "FLOAT8": TokenType.DOUBLE,
 904        "DOUBLE": TokenType.DOUBLE,
 905        "DOUBLE PRECISION": TokenType.DOUBLE,
 906        "JSON": TokenType.JSON,
 907        "JSONB": TokenType.JSONB,
 908        "CHAR": TokenType.CHAR,
 909        "CHARACTER": TokenType.CHAR,
 910        "NCHAR": TokenType.NCHAR,
 911        "VARCHAR": TokenType.VARCHAR,
 912        "VARCHAR2": TokenType.VARCHAR,
 913        "NVARCHAR": TokenType.NVARCHAR,
 914        "NVARCHAR2": TokenType.NVARCHAR,
 915        "BPCHAR": TokenType.BPCHAR,
 916        "STR": TokenType.TEXT,
 917        "STRING": TokenType.TEXT,
 918        "TEXT": TokenType.TEXT,
 919        "LONGTEXT": TokenType.LONGTEXT,
 920        "MEDIUMTEXT": TokenType.MEDIUMTEXT,
 921        "TINYTEXT": TokenType.TINYTEXT,
 922        "CLOB": TokenType.TEXT,
 923        "LONGVARCHAR": TokenType.TEXT,
 924        "BINARY": TokenType.BINARY,
 925        "BLOB": TokenType.VARBINARY,
 926        "LONGBLOB": TokenType.LONGBLOB,
 927        "MEDIUMBLOB": TokenType.MEDIUMBLOB,
 928        "TINYBLOB": TokenType.TINYBLOB,
 929        "BYTEA": TokenType.VARBINARY,
 930        "VARBINARY": TokenType.VARBINARY,
 931        "TIME": TokenType.TIME,
 932        "TIMETZ": TokenType.TIMETZ,
 933        "TIMESTAMP": TokenType.TIMESTAMP,
 934        "TIMESTAMPTZ": TokenType.TIMESTAMPTZ,
 935        "TIMESTAMPLTZ": TokenType.TIMESTAMPLTZ,
 936        "TIMESTAMP_LTZ": TokenType.TIMESTAMPLTZ,
 937        "TIMESTAMPNTZ": TokenType.TIMESTAMPNTZ,
 938        "TIMESTAMP_NTZ": TokenType.TIMESTAMPNTZ,
 939        "DATE": TokenType.DATE,
 940        "DATETIME": TokenType.DATETIME,
 941        "INT4RANGE": TokenType.INT4RANGE,
 942        "INT4MULTIRANGE": TokenType.INT4MULTIRANGE,
 943        "INT8RANGE": TokenType.INT8RANGE,
 944        "INT8MULTIRANGE": TokenType.INT8MULTIRANGE,
 945        "NUMRANGE": TokenType.NUMRANGE,
 946        "NUMMULTIRANGE": TokenType.NUMMULTIRANGE,
 947        "TSRANGE": TokenType.TSRANGE,
 948        "TSMULTIRANGE": TokenType.TSMULTIRANGE,
 949        "TSTZRANGE": TokenType.TSTZRANGE,
 950        "TSTZMULTIRANGE": TokenType.TSTZMULTIRANGE,
 951        "DATERANGE": TokenType.DATERANGE,
 952        "DATEMULTIRANGE": TokenType.DATEMULTIRANGE,
 953        "UNIQUE": TokenType.UNIQUE,
 954        "VECTOR": TokenType.VECTOR,
 955        "STRUCT": TokenType.STRUCT,
 956        "SEQUENCE": TokenType.SEQUENCE,
 957        "VARIANT": TokenType.VARIANT,
 958        "ALTER": TokenType.ALTER,
 959        "ANALYZE": TokenType.ANALYZE,
 960        "CALL": TokenType.COMMAND,
 961        "COMMENT": TokenType.COMMENT,
 962        "EXPLAIN": TokenType.COMMAND,
 963        "GRANT": TokenType.GRANT,
 964        "OPTIMIZE": TokenType.COMMAND,
 965        "PREPARE": TokenType.COMMAND,
 966        "VACUUM": TokenType.COMMAND,
 967        "USER-DEFINED": TokenType.USERDEFINED,
 968        "FOR VERSION": TokenType.VERSION_SNAPSHOT,
 969        "FOR TIMESTAMP": TokenType.TIMESTAMP_SNAPSHOT,
 970    }
 971
 972    WHITE_SPACE: t.Dict[t.Optional[str], TokenType] = {
 973        " ": TokenType.SPACE,
 974        "\t": TokenType.SPACE,
 975        "\n": TokenType.BREAK,
 976        "\r": TokenType.BREAK,
 977    }
 978
 979    COMMANDS = {
 980        TokenType.COMMAND,
 981        TokenType.EXECUTE,
 982        TokenType.FETCH,
 983        TokenType.SHOW,
 984        TokenType.RENAME,
 985    }
 986
 987    COMMAND_PREFIX_TOKENS = {TokenType.SEMICOLON, TokenType.BEGIN}
 988
 989    # Handle numeric literals like in hive (3L = BIGINT)
 990    NUMERIC_LITERALS: t.Dict[str, str] = {}
 991
 992    COMMENTS = ["--", ("/*", "*/")]
 993
 994    __slots__ = (
 995        "sql",
 996        "size",
 997        "tokens",
 998        "dialect",
 999        "use_rs_tokenizer",
1000        "_start",
1001        "_current",
1002        "_line",
1003        "_col",
1004        "_comments",
1005        "_char",
1006        "_end",
1007        "_peek",
1008        "_prev_token_line",
1009        "_rs_dialect_settings",
1010    )
1011
1012    def __init__(
1013        self, dialect: DialectType = None, use_rs_tokenizer: t.Optional[bool] = None
1014    ) -> None:
1015        from sqlglot.dialects import Dialect
1016
1017        self.dialect = Dialect.get_or_raise(dialect)
1018
1019        # initialize `use_rs_tokenizer`, and allow it to be overwritten per Tokenizer instance
1020        self.use_rs_tokenizer = (
1021            use_rs_tokenizer if use_rs_tokenizer is not None else USE_RS_TOKENIZER
1022        )
1023
1024        if self.use_rs_tokenizer:
1025            self._rs_dialect_settings = RsTokenizerDialectSettings(
1026                unescaped_sequences=self.dialect.UNESCAPED_SEQUENCES,
1027                identifiers_can_start_with_digit=self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT,
1028                numbers_can_be_underscore_separated=self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED,
1029            )
1030
1031        self.reset()
1032
1033    def reset(self) -> None:
1034        self.sql = ""
1035        self.size = 0
1036        self.tokens: t.List[Token] = []
1037        self._start = 0
1038        self._current = 0
1039        self._line = 1
1040        self._col = 0
1041        self._comments: t.List[str] = []
1042
1043        self._char = ""
1044        self._end = False
1045        self._peek = ""
1046        self._prev_token_line = -1
1047
1048    def tokenize(self, sql: str) -> t.List[Token]:
1049        """Returns a list of tokens corresponding to the SQL string `sql`."""
1050        if self.use_rs_tokenizer:
1051            return self.tokenize_rs(sql)
1052
1053        self.reset()
1054        self.sql = sql
1055        self.size = len(sql)
1056
1057        try:
1058            self._scan()
1059        except Exception as e:
1060            start = max(self._current - 50, 0)
1061            end = min(self._current + 50, self.size - 1)
1062            context = self.sql[start:end]
1063            raise TokenError(f"Error tokenizing '{context}'") from e
1064
1065        return self.tokens
1066
1067    def _scan(self, until: t.Optional[t.Callable] = None) -> None:
1068        while self.size and not self._end:
1069            current = self._current
1070
1071            # Skip spaces here rather than iteratively calling advance() for performance reasons
1072            while current < self.size:
1073                char = self.sql[current]
1074
1075                if char.isspace() and (char == " " or char == "\t"):
1076                    current += 1
1077                else:
1078                    break
1079
1080            offset = current - self._current if current > self._current else 1
1081
1082            self._start = current
1083            self._advance(offset)
1084
1085            if not self._char.isspace():
1086                if self._char.isdigit():
1087                    self._scan_number()
1088                elif self._char in self._IDENTIFIERS:
1089                    self._scan_identifier(self._IDENTIFIERS[self._char])
1090                else:
1091                    self._scan_keywords()
1092
1093            if until and until():
1094                break
1095
1096        if self.tokens and self._comments:
1097            self.tokens[-1].comments.extend(self._comments)
1098
1099    def _chars(self, size: int) -> str:
1100        if size == 1:
1101            return self._char
1102
1103        start = self._current - 1
1104        end = start + size
1105
1106        return self.sql[start:end] if end <= self.size else ""
1107
1108    def _advance(self, i: int = 1, alnum: bool = False) -> None:
1109        if self.WHITE_SPACE.get(self._char) is TokenType.BREAK:
1110            # Ensures we don't count an extra line if we get a \r\n line break sequence
1111            if not (self._char == "\r" and self._peek == "\n"):
1112                self._col = 1
1113                self._line += 1
1114        else:
1115            self._col += i
1116
1117        self._current += i
1118        self._end = self._current >= self.size
1119        self._char = self.sql[self._current - 1]
1120        self._peek = "" if self._end else self.sql[self._current]
1121
1122        if alnum and self._char.isalnum():
1123            # Here we use local variables instead of attributes for better performance
1124            _col = self._col
1125            _current = self._current
1126            _end = self._end
1127            _peek = self._peek
1128
1129            while _peek.isalnum():
1130                _col += 1
1131                _current += 1
1132                _end = _current >= self.size
1133                _peek = "" if _end else self.sql[_current]
1134
1135            self._col = _col
1136            self._current = _current
1137            self._end = _end
1138            self._peek = _peek
1139            self._char = self.sql[_current - 1]
1140
1141    @property
1142    def _text(self) -> str:
1143        return self.sql[self._start : self._current]
1144
1145    def _add(self, token_type: TokenType, text: t.Optional[str] = None) -> None:
1146        self._prev_token_line = self._line
1147
1148        if self._comments and token_type == TokenType.SEMICOLON and self.tokens:
1149            self.tokens[-1].comments.extend(self._comments)
1150            self._comments = []
1151
1152        self.tokens.append(
1153            Token(
1154                token_type,
1155                text=self._text if text is None else text,
1156                line=self._line,
1157                col=self._col,
1158                start=self._start,
1159                end=self._current - 1,
1160                comments=self._comments,
1161            )
1162        )
1163        self._comments = []
1164
1165        # If we have either a semicolon or a begin token before the command's token, we'll parse
1166        # whatever follows the command's token as a string
1167        if (
1168            token_type in self.COMMANDS
1169            and self._peek != ";"
1170            and (len(self.tokens) == 1 or self.tokens[-2].token_type in self.COMMAND_PREFIX_TOKENS)
1171        ):
1172            start = self._current
1173            tokens = len(self.tokens)
1174            self._scan(lambda: self._peek == ";")
1175            self.tokens = self.tokens[:tokens]
1176            text = self.sql[start : self._current].strip()
1177            if text:
1178                self._add(TokenType.STRING, text)
1179
1180    def _scan_keywords(self) -> None:
1181        size = 0
1182        word = None
1183        chars = self._text
1184        char = chars
1185        prev_space = False
1186        skip = False
1187        trie = self._KEYWORD_TRIE
1188        single_token = char in self.SINGLE_TOKENS
1189
1190        while chars:
1191            if skip:
1192                result = TrieResult.PREFIX
1193            else:
1194                result, trie = in_trie(trie, char.upper())
1195
1196            if result == TrieResult.FAILED:
1197                break
1198            if result == TrieResult.EXISTS:
1199                word = chars
1200
1201            end = self._current + size
1202            size += 1
1203
1204            if end < self.size:
1205                char = self.sql[end]
1206                single_token = single_token or char in self.SINGLE_TOKENS
1207                is_space = char.isspace()
1208
1209                if not is_space or not prev_space:
1210                    if is_space:
1211                        char = " "
1212                    chars += char
1213                    prev_space = is_space
1214                    skip = False
1215                else:
1216                    skip = True
1217            else:
1218                char = ""
1219                break
1220
1221        if word:
1222            if self._scan_string(word):
1223                return
1224            if self._scan_comment(word):
1225                return
1226            if prev_space or single_token or not char:
1227                self._advance(size - 1)
1228                word = word.upper()
1229                self._add(self.KEYWORDS[word], text=word)
1230                return
1231
1232        if self._char in self.SINGLE_TOKENS:
1233            self._add(self.SINGLE_TOKENS[self._char], text=self._char)
1234            return
1235
1236        self._scan_var()
1237
1238    def _scan_comment(self, comment_start: str) -> bool:
1239        if comment_start not in self._COMMENTS:
1240            return False
1241
1242        comment_start_line = self._line
1243        comment_start_size = len(comment_start)
1244        comment_end = self._COMMENTS[comment_start]
1245
1246        if comment_end:
1247            # Skip the comment's start delimiter
1248            self._advance(comment_start_size)
1249
1250            comment_count = 1
1251            comment_end_size = len(comment_end)
1252
1253            while not self._end:
1254                if self._chars(comment_end_size) == comment_end:
1255                    comment_count -= 1
1256                    if not comment_count:
1257                        break
1258
1259                self._advance(alnum=True)
1260
1261                # Nested comments are allowed by some dialects, e.g. databricks, duckdb, postgres
1262                if (
1263                    self.NESTED_COMMENTS
1264                    and not self._end
1265                    and self._chars(comment_end_size) == comment_start
1266                ):
1267                    self._advance(comment_start_size)
1268                    comment_count += 1
1269
1270            self._comments.append(self._text[comment_start_size : -comment_end_size + 1])
1271            self._advance(comment_end_size - 1)
1272        else:
1273            while not self._end and self.WHITE_SPACE.get(self._peek) is not TokenType.BREAK:
1274                self._advance(alnum=True)
1275            self._comments.append(self._text[comment_start_size:])
1276
1277        if (
1278            comment_start == self.HINT_START
1279            and self.tokens
1280            and self.tokens[-1].token_type in self.TOKENS_PRECEDING_HINT
1281        ):
1282            self._add(TokenType.HINT)
1283
1284        # Leading comment is attached to the succeeding token, whilst trailing comment to the preceding.
1285        # Multiple consecutive comments are preserved by appending them to the current comments list.
1286        if comment_start_line == self._prev_token_line:
1287            self.tokens[-1].comments.extend(self._comments)
1288            self._comments = []
1289            self._prev_token_line = self._line
1290
1291        return True
1292
1293    def _scan_number(self) -> None:
1294        if self._char == "0":
1295            peek = self._peek.upper()
1296            if peek == "B":
1297                return self._scan_bits() if self.BIT_STRINGS else self._add(TokenType.NUMBER)
1298            elif peek == "X":
1299                return self._scan_hex() if self.HEX_STRINGS else self._add(TokenType.NUMBER)
1300
1301        decimal = False
1302        scientific = 0
1303
1304        while True:
1305            if self._peek.isdigit():
1306                self._advance()
1307            elif self._peek == "." and not decimal:
1308                if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER:
1309                    return self._add(TokenType.NUMBER)
1310                decimal = True
1311                self._advance()
1312            elif self._peek in ("-", "+") and scientific == 1:
1313                scientific += 1
1314                self._advance()
1315            elif self._peek.upper() == "E" and not scientific:
1316                scientific += 1
1317                self._advance()
1318            elif self._peek.isidentifier():
1319                number_text = self._text
1320                literal = ""
1321
1322                while self._peek.strip() and self._peek not in self.SINGLE_TOKENS:
1323                    literal += self._peek
1324                    self._advance()
1325
1326                token_type = self.KEYWORDS.get(self.NUMERIC_LITERALS.get(literal.upper(), ""))
1327
1328                if token_type:
1329                    self._add(TokenType.NUMBER, number_text)
1330                    self._add(TokenType.DCOLON, "::")
1331                    return self._add(token_type, literal)
1332                else:
1333                    replaced = literal.replace("_", "")
1334                    if self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED and replaced.isdigit():
1335                        return self._add(TokenType.NUMBER, number_text + replaced)
1336                    if self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT:
1337                        return self._add(TokenType.VAR)
1338
1339                self._advance(-len(literal))
1340                return self._add(TokenType.NUMBER, number_text)
1341            else:
1342                return self._add(TokenType.NUMBER)
1343
1344    def _scan_bits(self) -> None:
1345        self._advance()
1346        value = self._extract_value()
1347        try:
1348            # If `value` can't be converted to a binary, fallback to tokenizing it as an identifier
1349            int(value, 2)
1350            self._add(TokenType.BIT_STRING, value[2:])  # Drop the 0b
1351        except ValueError:
1352            self._add(TokenType.IDENTIFIER)
1353
1354    def _scan_hex(self) -> None:
1355        self._advance()
1356        value = self._extract_value()
1357        try:
1358            # If `value` can't be converted to a hex, fallback to tokenizing it as an identifier
1359            int(value, 16)
1360            self._add(TokenType.HEX_STRING, value[2:])  # Drop the 0x
1361        except ValueError:
1362            self._add(TokenType.IDENTIFIER)
1363
1364    def _extract_value(self) -> str:
1365        while True:
1366            char = self._peek.strip()
1367            if char and char not in self.SINGLE_TOKENS:
1368                self._advance(alnum=True)
1369            else:
1370                break
1371
1372        return self._text
1373
1374    def _scan_string(self, start: str) -> bool:
1375        base = None
1376        token_type = TokenType.STRING
1377
1378        if start in self._QUOTES:
1379            end = self._QUOTES[start]
1380        elif start in self._FORMAT_STRINGS:
1381            end, token_type = self._FORMAT_STRINGS[start]
1382
1383            if token_type == TokenType.HEX_STRING:
1384                base = 16
1385            elif token_type == TokenType.BIT_STRING:
1386                base = 2
1387            elif token_type == TokenType.HEREDOC_STRING:
1388                self._advance()
1389
1390                if self._char == end:
1391                    tag = ""
1392                else:
1393                    tag = self._extract_string(
1394                        end,
1395                        raw_string=True,
1396                        raise_unmatched=not self.HEREDOC_TAG_IS_IDENTIFIER,
1397                    )
1398
1399                if tag and self.HEREDOC_TAG_IS_IDENTIFIER and (self._end or not tag.isidentifier()):
1400                    if not self._end:
1401                        self._advance(-1)
1402
1403                    self._advance(-len(tag))
1404                    self._add(self.HEREDOC_STRING_ALTERNATIVE)
1405                    return True
1406
1407                end = f"{start}{tag}{end}"
1408        else:
1409            return False
1410
1411        self._advance(len(start))
1412        text = self._extract_string(end, raw_string=token_type == TokenType.RAW_STRING)
1413
1414        if base:
1415            try:
1416                int(text, base)
1417            except Exception:
1418                raise TokenError(
1419                    f"Numeric string contains invalid characters from {self._line}:{self._start}"
1420                )
1421
1422        self._add(token_type, text)
1423        return True
1424
1425    def _scan_identifier(self, identifier_end: str) -> None:
1426        self._advance()
1427        text = self._extract_string(
1428            identifier_end, escapes=self._IDENTIFIER_ESCAPES | {identifier_end}
1429        )
1430        self._add(TokenType.IDENTIFIER, text)
1431
1432    def _scan_var(self) -> None:
1433        while True:
1434            char = self._peek.strip()
1435            if char and (char in self.VAR_SINGLE_TOKENS or char not in self.SINGLE_TOKENS):
1436                self._advance(alnum=True)
1437            else:
1438                break
1439
1440        self._add(
1441            TokenType.VAR
1442            if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER
1443            else self.KEYWORDS.get(self._text.upper(), TokenType.VAR)
1444        )
1445
1446    def _extract_string(
1447        self,
1448        delimiter: str,
1449        escapes: t.Optional[t.Set[str]] = None,
1450        raw_string: bool = False,
1451        raise_unmatched: bool = True,
1452    ) -> str:
1453        text = ""
1454        delim_size = len(delimiter)
1455        escapes = self._STRING_ESCAPES if escapes is None else escapes
1456
1457        while True:
1458            if (
1459                not raw_string
1460                and self.dialect.UNESCAPED_SEQUENCES
1461                and self._peek
1462                and self._char in self.STRING_ESCAPES
1463            ):
1464                unescaped_sequence = self.dialect.UNESCAPED_SEQUENCES.get(self._char + self._peek)
1465                if unescaped_sequence:
1466                    self._advance(2)
1467                    text += unescaped_sequence
1468                    continue
1469            if (
1470                (self.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS or not raw_string)
1471                and self._char in escapes
1472                and (self._peek == delimiter or self._peek in escapes)
1473                and (self._char not in self._QUOTES or self._char == self._peek)
1474            ):
1475                if self._peek == delimiter:
1476                    text += self._peek
1477                else:
1478                    text += self._char + self._peek
1479
1480                if self._current + 1 < self.size:
1481                    self._advance(2)
1482                else:
1483                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._current}")
1484            else:
1485                if self._chars(delim_size) == delimiter:
1486                    if delim_size > 1:
1487                        self._advance(delim_size - 1)
1488                    break
1489
1490                if self._end:
1491                    if not raise_unmatched:
1492                        return text + self._char
1493
1494                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._start}")
1495
1496                current = self._current - 1
1497                self._advance(alnum=True)
1498                text += self.sql[current : self._current - 1]
1499
1500        return text
1501
1502    def tokenize_rs(self, sql: str) -> t.List[Token]:
1503        if not self._RS_TOKENIZER:
1504            raise SqlglotError("Rust tokenizer is not available")
1505
1506        try:
1507            tokens = self._RS_TOKENIZER.tokenize(sql, self._rs_dialect_settings)
1508            for token in tokens:
1509                token.token_type = _ALL_TOKEN_TYPES[token.token_type_index]
1510            return tokens
1511        except Exception as e:
1512            raise TokenError(str(e))