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