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