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