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