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