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    VECTOR = auto()
 206
 207    # keywords
 208    ALIAS = auto()
 209    ALTER = auto()
 210    ALWAYS = auto()
 211    ALL = auto()
 212    ANTI = auto()
 213    ANY = auto()
 214    APPLY = auto()
 215    ARRAY = auto()
 216    ASC = auto()
 217    ASOF = auto()
 218    AUTO_INCREMENT = auto()
 219    BEGIN = auto()
 220    BETWEEN = auto()
 221    CACHE = auto()
 222    CASE = auto()
 223    CHARACTER_SET = auto()
 224    CLUSTER_BY = auto()
 225    COLLATE = auto()
 226    COMMAND = auto()
 227    COMMENT = auto()
 228    COMMIT = auto()
 229    CONNECT_BY = auto()
 230    CONSTRAINT = auto()
 231    COPY = auto()
 232    CREATE = auto()
 233    CROSS = auto()
 234    CUBE = auto()
 235    CURRENT_DATE = auto()
 236    CURRENT_DATETIME = auto()
 237    CURRENT_TIME = auto()
 238    CURRENT_TIMESTAMP = auto()
 239    CURRENT_USER = auto()
 240    DECLARE = auto()
 241    DEFAULT = auto()
 242    DELETE = auto()
 243    DESC = auto()
 244    DESCRIBE = auto()
 245    DICTIONARY = auto()
 246    DISTINCT = auto()
 247    DISTRIBUTE_BY = auto()
 248    DIV = auto()
 249    DROP = auto()
 250    ELSE = auto()
 251    END = auto()
 252    ESCAPE = auto()
 253    EXCEPT = auto()
 254    EXECUTE = auto()
 255    EXISTS = auto()
 256    FALSE = auto()
 257    FETCH = auto()
 258    FILTER = auto()
 259    FINAL = auto()
 260    FIRST = auto()
 261    FOR = auto()
 262    FORCE = auto()
 263    FOREIGN_KEY = auto()
 264    FORMAT = auto()
 265    FROM = auto()
 266    FULL = auto()
 267    FUNCTION = auto()
 268    GLOB = auto()
 269    GLOBAL = auto()
 270    GROUP_BY = auto()
 271    GROUPING_SETS = auto()
 272    HAVING = auto()
 273    HINT = auto()
 274    IGNORE = auto()
 275    ILIKE = auto()
 276    ILIKE_ANY = auto()
 277    IN = auto()
 278    INDEX = auto()
 279    INNER = auto()
 280    INSERT = auto()
 281    INTERSECT = auto()
 282    INTERVAL = auto()
 283    INTO = auto()
 284    INTRODUCER = auto()
 285    IRLIKE = auto()
 286    IS = auto()
 287    ISNULL = auto()
 288    JOIN = auto()
 289    JOIN_MARKER = auto()
 290    KEEP = auto()
 291    KEY = auto()
 292    KILL = auto()
 293    LANGUAGE = auto()
 294    LATERAL = auto()
 295    LEFT = auto()
 296    LIKE = auto()
 297    LIKE_ANY = auto()
 298    LIMIT = auto()
 299    LIST = auto()
 300    LOAD = auto()
 301    LOCK = auto()
 302    MAP = auto()
 303    MATCH_CONDITION = auto()
 304    MATCH_RECOGNIZE = auto()
 305    MEMBER_OF = auto()
 306    MERGE = auto()
 307    MOD = auto()
 308    MODEL = auto()
 309    NATURAL = auto()
 310    NEXT = auto()
 311    NOTNULL = auto()
 312    NULL = auto()
 313    OBJECT_IDENTIFIER = auto()
 314    OFFSET = auto()
 315    ON = auto()
 316    ONLY = auto()
 317    OPERATOR = auto()
 318    ORDER_BY = auto()
 319    ORDER_SIBLINGS_BY = auto()
 320    ORDERED = auto()
 321    ORDINALITY = auto()
 322    OUTER = auto()
 323    OVER = auto()
 324    OVERLAPS = auto()
 325    OVERWRITE = auto()
 326    PARTITION = auto()
 327    PARTITION_BY = auto()
 328    PERCENT = auto()
 329    PIVOT = auto()
 330    PLACEHOLDER = auto()
 331    POSITIONAL = auto()
 332    PRAGMA = auto()
 333    PREWHERE = auto()
 334    PRIMARY_KEY = auto()
 335    PROCEDURE = auto()
 336    PROPERTIES = auto()
 337    PSEUDO_TYPE = auto()
 338    QUALIFY = auto()
 339    QUOTE = auto()
 340    RANGE = auto()
 341    RECURSIVE = auto()
 342    REFRESH = auto()
 343    RENAME = auto()
 344    REPLACE = auto()
 345    RETURNING = auto()
 346    REFERENCES = auto()
 347    RIGHT = auto()
 348    RLIKE = auto()
 349    ROLLBACK = auto()
 350    ROLLUP = auto()
 351    ROW = auto()
 352    ROWS = auto()
 353    SELECT = auto()
 354    SEMI = auto()
 355    SEPARATOR = auto()
 356    SEQUENCE = auto()
 357    SERDE_PROPERTIES = auto()
 358    SET = auto()
 359    SETTINGS = auto()
 360    SHOW = auto()
 361    SIMILAR_TO = auto()
 362    SOME = auto()
 363    SORT_BY = auto()
 364    START_WITH = auto()
 365    STORAGE_INTEGRATION = auto()
 366    STRAIGHT_JOIN = auto()
 367    STRUCT = auto()
 368    SUMMARIZE = auto()
 369    TABLE_SAMPLE = auto()
 370    TAG = auto()
 371    TEMPORARY = auto()
 372    TOP = auto()
 373    THEN = auto()
 374    TRUE = auto()
 375    TRUNCATE = auto()
 376    UNCACHE = auto()
 377    UNION = auto()
 378    UNNEST = auto()
 379    UNPIVOT = auto()
 380    UPDATE = auto()
 381    USE = auto()
 382    USING = auto()
 383    VALUES = auto()
 384    VIEW = auto()
 385    VOLATILE = auto()
 386    WHEN = auto()
 387    WHERE = auto()
 388    WINDOW = auto()
 389    WITH = auto()
 390    UNIQUE = auto()
 391    VERSION_SNAPSHOT = auto()
 392    TIMESTAMP_SNAPSHOT = auto()
 393    OPTION = auto()
 394
 395
 396_ALL_TOKEN_TYPES = list(TokenType)
 397_TOKEN_TYPE_TO_INDEX = {token_type: i for i, token_type in enumerate(_ALL_TOKEN_TYPES)}
 398
 399
 400class Token:
 401    __slots__ = ("token_type", "text", "line", "col", "start", "end", "comments")
 402
 403    @classmethod
 404    def number(cls, number: int) -> Token:
 405        """Returns a NUMBER token with `number` as its text."""
 406        return cls(TokenType.NUMBER, str(number))
 407
 408    @classmethod
 409    def string(cls, string: str) -> Token:
 410        """Returns a STRING token with `string` as its text."""
 411        return cls(TokenType.STRING, string)
 412
 413    @classmethod
 414    def identifier(cls, identifier: str) -> Token:
 415        """Returns an IDENTIFIER token with `identifier` as its text."""
 416        return cls(TokenType.IDENTIFIER, identifier)
 417
 418    @classmethod
 419    def var(cls, var: str) -> Token:
 420        """Returns an VAR token with `var` as its text."""
 421        return cls(TokenType.VAR, var)
 422
 423    def __init__(
 424        self,
 425        token_type: TokenType,
 426        text: str,
 427        line: int = 1,
 428        col: int = 1,
 429        start: int = 0,
 430        end: int = 0,
 431        comments: t.Optional[t.List[str]] = None,
 432    ) -> None:
 433        """Token initializer.
 434
 435        Args:
 436            token_type: The TokenType Enum.
 437            text: The text of the token.
 438            line: The line that the token ends on.
 439            col: The column that the token ends on.
 440            start: The start index of the token.
 441            end: The ending index of the token.
 442            comments: The comments to attach to the token.
 443        """
 444        self.token_type = token_type
 445        self.text = text
 446        self.line = line
 447        self.col = col
 448        self.start = start
 449        self.end = end
 450        self.comments = [] if comments is None else comments
 451
 452    def __repr__(self) -> str:
 453        attributes = ", ".join(f"{k}: {getattr(self, k)}" for k in self.__slots__)
 454        return f"<Token {attributes}>"
 455
 456
 457class _Tokenizer(type):
 458    def __new__(cls, clsname, bases, attrs):
 459        klass = super().__new__(cls, clsname, bases, attrs)
 460
 461        def _convert_quotes(arr: t.List[str | t.Tuple[str, str]]) -> t.Dict[str, str]:
 462            return dict(
 463                (item, item) if isinstance(item, str) else (item[0], item[1]) for item in arr
 464            )
 465
 466        def _quotes_to_format(
 467            token_type: TokenType, arr: t.List[str | t.Tuple[str, str]]
 468        ) -> t.Dict[str, t.Tuple[str, TokenType]]:
 469            return {k: (v, token_type) for k, v in _convert_quotes(arr).items()}
 470
 471        klass._QUOTES = _convert_quotes(klass.QUOTES)
 472        klass._IDENTIFIERS = _convert_quotes(klass.IDENTIFIERS)
 473
 474        klass._FORMAT_STRINGS = {
 475            **{
 476                p + s: (e, TokenType.NATIONAL_STRING)
 477                for s, e in klass._QUOTES.items()
 478                for p in ("n", "N")
 479            },
 480            **_quotes_to_format(TokenType.BIT_STRING, klass.BIT_STRINGS),
 481            **_quotes_to_format(TokenType.BYTE_STRING, klass.BYTE_STRINGS),
 482            **_quotes_to_format(TokenType.HEX_STRING, klass.HEX_STRINGS),
 483            **_quotes_to_format(TokenType.RAW_STRING, klass.RAW_STRINGS),
 484            **_quotes_to_format(TokenType.HEREDOC_STRING, klass.HEREDOC_STRINGS),
 485            **_quotes_to_format(TokenType.UNICODE_STRING, klass.UNICODE_STRINGS),
 486        }
 487
 488        klass._STRING_ESCAPES = set(klass.STRING_ESCAPES)
 489        klass._IDENTIFIER_ESCAPES = set(klass.IDENTIFIER_ESCAPES)
 490        klass._COMMENTS = {
 491            **dict(
 492                (comment, None) if isinstance(comment, str) else (comment[0], comment[1])
 493                for comment in klass.COMMENTS
 494            ),
 495            "{#": "#}",  # Ensure Jinja comments are tokenized correctly in all dialects
 496        }
 497
 498        klass._KEYWORD_TRIE = new_trie(
 499            key.upper()
 500            for key in (
 501                *klass.KEYWORDS,
 502                *klass._COMMENTS,
 503                *klass._QUOTES,
 504                *klass._FORMAT_STRINGS,
 505            )
 506            if " " in key or any(single in key for single in klass.SINGLE_TOKENS)
 507        )
 508
 509        if USE_RS_TOKENIZER:
 510            settings = RsTokenizerSettings(
 511                white_space={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.WHITE_SPACE.items()},
 512                single_tokens={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.SINGLE_TOKENS.items()},
 513                keywords={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.KEYWORDS.items()},
 514                numeric_literals=klass.NUMERIC_LITERALS,
 515                identifiers=klass._IDENTIFIERS,
 516                identifier_escapes=klass._IDENTIFIER_ESCAPES,
 517                string_escapes=klass._STRING_ESCAPES,
 518                quotes=klass._QUOTES,
 519                format_strings={
 520                    k: (v1, _TOKEN_TYPE_TO_INDEX[v2])
 521                    for k, (v1, v2) in klass._FORMAT_STRINGS.items()
 522                },
 523                has_bit_strings=bool(klass.BIT_STRINGS),
 524                has_hex_strings=bool(klass.HEX_STRINGS),
 525                comments=klass._COMMENTS,
 526                var_single_tokens=klass.VAR_SINGLE_TOKENS,
 527                commands={_TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMANDS},
 528                command_prefix_tokens={
 529                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMAND_PREFIX_TOKENS
 530                },
 531                heredoc_tag_is_identifier=klass.HEREDOC_TAG_IS_IDENTIFIER,
 532                string_escapes_allowed_in_raw_strings=klass.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS,
 533                nested_comments=klass.NESTED_COMMENTS,
 534            )
 535            token_types = RsTokenTypeSettings(
 536                bit_string=_TOKEN_TYPE_TO_INDEX[TokenType.BIT_STRING],
 537                break_=_TOKEN_TYPE_TO_INDEX[TokenType.BREAK],
 538                dcolon=_TOKEN_TYPE_TO_INDEX[TokenType.DCOLON],
 539                heredoc_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEREDOC_STRING],
 540                raw_string=_TOKEN_TYPE_TO_INDEX[TokenType.RAW_STRING],
 541                hex_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEX_STRING],
 542                identifier=_TOKEN_TYPE_TO_INDEX[TokenType.IDENTIFIER],
 543                number=_TOKEN_TYPE_TO_INDEX[TokenType.NUMBER],
 544                parameter=_TOKEN_TYPE_TO_INDEX[TokenType.PARAMETER],
 545                semicolon=_TOKEN_TYPE_TO_INDEX[TokenType.SEMICOLON],
 546                string=_TOKEN_TYPE_TO_INDEX[TokenType.STRING],
 547                var=_TOKEN_TYPE_TO_INDEX[TokenType.VAR],
 548                heredoc_string_alternative=_TOKEN_TYPE_TO_INDEX[klass.HEREDOC_STRING_ALTERNATIVE],
 549            )
 550            klass._RS_TOKENIZER = RsTokenizer(settings, token_types)
 551        else:
 552            klass._RS_TOKENIZER = None
 553
 554        return klass
 555
 556
 557class Tokenizer(metaclass=_Tokenizer):
 558    SINGLE_TOKENS = {
 559        "(": TokenType.L_PAREN,
 560        ")": TokenType.R_PAREN,
 561        "[": TokenType.L_BRACKET,
 562        "]": TokenType.R_BRACKET,
 563        "{": TokenType.L_BRACE,
 564        "}": TokenType.R_BRACE,
 565        "&": TokenType.AMP,
 566        "^": TokenType.CARET,
 567        ":": TokenType.COLON,
 568        ",": TokenType.COMMA,
 569        ".": TokenType.DOT,
 570        "-": TokenType.DASH,
 571        "=": TokenType.EQ,
 572        ">": TokenType.GT,
 573        "<": TokenType.LT,
 574        "%": TokenType.MOD,
 575        "!": TokenType.NOT,
 576        "|": TokenType.PIPE,
 577        "+": TokenType.PLUS,
 578        ";": TokenType.SEMICOLON,
 579        "/": TokenType.SLASH,
 580        "\\": TokenType.BACKSLASH,
 581        "*": TokenType.STAR,
 582        "~": TokenType.TILDA,
 583        "?": TokenType.PLACEHOLDER,
 584        "@": TokenType.PARAMETER,
 585        "#": TokenType.HASH,
 586        # Used for breaking a var like x'y' but nothing else the token type doesn't matter
 587        "'": TokenType.UNKNOWN,
 588        "`": TokenType.UNKNOWN,
 589        '"': TokenType.UNKNOWN,
 590    }
 591
 592    BIT_STRINGS: t.List[str | t.Tuple[str, str]] = []
 593    BYTE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 594    HEX_STRINGS: t.List[str | t.Tuple[str, str]] = []
 595    RAW_STRINGS: t.List[str | t.Tuple[str, str]] = []
 596    HEREDOC_STRINGS: t.List[str | t.Tuple[str, str]] = []
 597    UNICODE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 598    IDENTIFIERS: t.List[str | t.Tuple[str, str]] = ['"']
 599    IDENTIFIER_ESCAPES = ['"']
 600    QUOTES: t.List[t.Tuple[str, str] | str] = ["'"]
 601    STRING_ESCAPES = ["'"]
 602    VAR_SINGLE_TOKENS: t.Set[str] = set()
 603
 604    # Whether the heredoc tags follow the same lexical rules as unquoted identifiers
 605    HEREDOC_TAG_IS_IDENTIFIER = False
 606
 607    # Token that we'll generate as a fallback if the heredoc prefix doesn't correspond to a heredoc
 608    HEREDOC_STRING_ALTERNATIVE = TokenType.VAR
 609
 610    # Whether string escape characters function as such when placed within raw strings
 611    STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS = True
 612
 613    NESTED_COMMENTS = True
 614
 615    # Autofilled
 616    _COMMENTS: t.Dict[str, str] = {}
 617    _FORMAT_STRINGS: t.Dict[str, t.Tuple[str, TokenType]] = {}
 618    _IDENTIFIERS: t.Dict[str, str] = {}
 619    _IDENTIFIER_ESCAPES: t.Set[str] = set()
 620    _QUOTES: t.Dict[str, str] = {}
 621    _STRING_ESCAPES: t.Set[str] = set()
 622    _KEYWORD_TRIE: t.Dict = {}
 623    _RS_TOKENIZER: t.Optional[t.Any] = None
 624
 625    KEYWORDS: t.Dict[str, TokenType] = {
 626        **{f"{{%{postfix}": TokenType.BLOCK_START for postfix in ("", "+", "-")},
 627        **{f"{prefix}%}}": TokenType.BLOCK_END for prefix in ("", "+", "-")},
 628        **{f"{{{{{postfix}": TokenType.BLOCK_START for postfix in ("+", "-")},
 629        **{f"{prefix}}}}}": TokenType.BLOCK_END for prefix in ("+", "-")},
 630        "/*+": TokenType.HINT,
 631        "==": TokenType.EQ,
 632        "::": TokenType.DCOLON,
 633        "||": TokenType.DPIPE,
 634        ">=": TokenType.GTE,
 635        "<=": TokenType.LTE,
 636        "<>": TokenType.NEQ,
 637        "!=": TokenType.NEQ,
 638        ":=": TokenType.COLON_EQ,
 639        "<=>": TokenType.NULLSAFE_EQ,
 640        "->": TokenType.ARROW,
 641        "->>": TokenType.DARROW,
 642        "=>": TokenType.FARROW,
 643        "#>": TokenType.HASH_ARROW,
 644        "#>>": TokenType.DHASH_ARROW,
 645        "<->": TokenType.LR_ARROW,
 646        "&&": TokenType.DAMP,
 647        "??": TokenType.DQMARK,
 648        "ALL": TokenType.ALL,
 649        "ALWAYS": TokenType.ALWAYS,
 650        "AND": TokenType.AND,
 651        "ANTI": TokenType.ANTI,
 652        "ANY": TokenType.ANY,
 653        "ASC": TokenType.ASC,
 654        "AS": TokenType.ALIAS,
 655        "ASOF": TokenType.ASOF,
 656        "AUTOINCREMENT": TokenType.AUTO_INCREMENT,
 657        "AUTO_INCREMENT": TokenType.AUTO_INCREMENT,
 658        "BEGIN": TokenType.BEGIN,
 659        "BETWEEN": TokenType.BETWEEN,
 660        "CACHE": TokenType.CACHE,
 661        "UNCACHE": TokenType.UNCACHE,
 662        "CASE": TokenType.CASE,
 663        "CHARACTER SET": TokenType.CHARACTER_SET,
 664        "CLUSTER BY": TokenType.CLUSTER_BY,
 665        "COLLATE": TokenType.COLLATE,
 666        "COLUMN": TokenType.COLUMN,
 667        "COMMIT": TokenType.COMMIT,
 668        "CONNECT BY": TokenType.CONNECT_BY,
 669        "CONSTRAINT": TokenType.CONSTRAINT,
 670        "COPY": TokenType.COPY,
 671        "CREATE": TokenType.CREATE,
 672        "CROSS": TokenType.CROSS,
 673        "CUBE": TokenType.CUBE,
 674        "CURRENT_DATE": TokenType.CURRENT_DATE,
 675        "CURRENT_TIME": TokenType.CURRENT_TIME,
 676        "CURRENT_TIMESTAMP": TokenType.CURRENT_TIMESTAMP,
 677        "CURRENT_USER": TokenType.CURRENT_USER,
 678        "DATABASE": TokenType.DATABASE,
 679        "DEFAULT": TokenType.DEFAULT,
 680        "DELETE": TokenType.DELETE,
 681        "DESC": TokenType.DESC,
 682        "DESCRIBE": TokenType.DESCRIBE,
 683        "DISTINCT": TokenType.DISTINCT,
 684        "DISTRIBUTE BY": TokenType.DISTRIBUTE_BY,
 685        "DIV": TokenType.DIV,
 686        "DROP": TokenType.DROP,
 687        "ELSE": TokenType.ELSE,
 688        "END": TokenType.END,
 689        "ENUM": TokenType.ENUM,
 690        "ESCAPE": TokenType.ESCAPE,
 691        "EXCEPT": TokenType.EXCEPT,
 692        "EXECUTE": TokenType.EXECUTE,
 693        "EXISTS": TokenType.EXISTS,
 694        "FALSE": TokenType.FALSE,
 695        "FETCH": TokenType.FETCH,
 696        "FILTER": TokenType.FILTER,
 697        "FIRST": TokenType.FIRST,
 698        "FULL": TokenType.FULL,
 699        "FUNCTION": TokenType.FUNCTION,
 700        "FOR": TokenType.FOR,
 701        "FOREIGN KEY": TokenType.FOREIGN_KEY,
 702        "FORMAT": TokenType.FORMAT,
 703        "FROM": TokenType.FROM,
 704        "GEOGRAPHY": TokenType.GEOGRAPHY,
 705        "GEOMETRY": TokenType.GEOMETRY,
 706        "GLOB": TokenType.GLOB,
 707        "GROUP BY": TokenType.GROUP_BY,
 708        "GROUPING SETS": TokenType.GROUPING_SETS,
 709        "HAVING": TokenType.HAVING,
 710        "ILIKE": TokenType.ILIKE,
 711        "IN": TokenType.IN,
 712        "INDEX": TokenType.INDEX,
 713        "INET": TokenType.INET,
 714        "INNER": TokenType.INNER,
 715        "INSERT": TokenType.INSERT,
 716        "INTERVAL": TokenType.INTERVAL,
 717        "INTERSECT": TokenType.INTERSECT,
 718        "INTO": TokenType.INTO,
 719        "IS": TokenType.IS,
 720        "ISNULL": TokenType.ISNULL,
 721        "JOIN": TokenType.JOIN,
 722        "KEEP": TokenType.KEEP,
 723        "KILL": TokenType.KILL,
 724        "LATERAL": TokenType.LATERAL,
 725        "LEFT": TokenType.LEFT,
 726        "LIKE": TokenType.LIKE,
 727        "LIMIT": TokenType.LIMIT,
 728        "LOAD": TokenType.LOAD,
 729        "LOCK": TokenType.LOCK,
 730        "MERGE": TokenType.MERGE,
 731        "NATURAL": TokenType.NATURAL,
 732        "NEXT": TokenType.NEXT,
 733        "NOT": TokenType.NOT,
 734        "NOTNULL": TokenType.NOTNULL,
 735        "NULL": TokenType.NULL,
 736        "OBJECT": TokenType.OBJECT,
 737        "OFFSET": TokenType.OFFSET,
 738        "ON": TokenType.ON,
 739        "OR": TokenType.OR,
 740        "XOR": TokenType.XOR,
 741        "ORDER BY": TokenType.ORDER_BY,
 742        "ORDINALITY": TokenType.ORDINALITY,
 743        "OUTER": TokenType.OUTER,
 744        "OVER": TokenType.OVER,
 745        "OVERLAPS": TokenType.OVERLAPS,
 746        "OVERWRITE": TokenType.OVERWRITE,
 747        "PARTITION": TokenType.PARTITION,
 748        "PARTITION BY": TokenType.PARTITION_BY,
 749        "PARTITIONED BY": TokenType.PARTITION_BY,
 750        "PARTITIONED_BY": TokenType.PARTITION_BY,
 751        "PERCENT": TokenType.PERCENT,
 752        "PIVOT": TokenType.PIVOT,
 753        "PRAGMA": TokenType.PRAGMA,
 754        "PRIMARY KEY": TokenType.PRIMARY_KEY,
 755        "PROCEDURE": TokenType.PROCEDURE,
 756        "QUALIFY": TokenType.QUALIFY,
 757        "RANGE": TokenType.RANGE,
 758        "RECURSIVE": TokenType.RECURSIVE,
 759        "REGEXP": TokenType.RLIKE,
 760        "RENAME": TokenType.RENAME,
 761        "REPLACE": TokenType.REPLACE,
 762        "RETURNING": TokenType.RETURNING,
 763        "REFERENCES": TokenType.REFERENCES,
 764        "RIGHT": TokenType.RIGHT,
 765        "RLIKE": TokenType.RLIKE,
 766        "ROLLBACK": TokenType.ROLLBACK,
 767        "ROLLUP": TokenType.ROLLUP,
 768        "ROW": TokenType.ROW,
 769        "ROWS": TokenType.ROWS,
 770        "SCHEMA": TokenType.SCHEMA,
 771        "SELECT": TokenType.SELECT,
 772        "SEMI": TokenType.SEMI,
 773        "SET": TokenType.SET,
 774        "SETTINGS": TokenType.SETTINGS,
 775        "SHOW": TokenType.SHOW,
 776        "SIMILAR TO": TokenType.SIMILAR_TO,
 777        "SOME": TokenType.SOME,
 778        "SORT BY": TokenType.SORT_BY,
 779        "START WITH": TokenType.START_WITH,
 780        "STRAIGHT_JOIN": TokenType.STRAIGHT_JOIN,
 781        "TABLE": TokenType.TABLE,
 782        "TABLESAMPLE": TokenType.TABLE_SAMPLE,
 783        "TEMP": TokenType.TEMPORARY,
 784        "TEMPORARY": TokenType.TEMPORARY,
 785        "THEN": TokenType.THEN,
 786        "TRUE": TokenType.TRUE,
 787        "TRUNCATE": TokenType.TRUNCATE,
 788        "UNION": TokenType.UNION,
 789        "UNKNOWN": TokenType.UNKNOWN,
 790        "UNNEST": TokenType.UNNEST,
 791        "UNPIVOT": TokenType.UNPIVOT,
 792        "UPDATE": TokenType.UPDATE,
 793        "USE": TokenType.USE,
 794        "USING": TokenType.USING,
 795        "UUID": TokenType.UUID,
 796        "VALUES": TokenType.VALUES,
 797        "VIEW": TokenType.VIEW,
 798        "VOLATILE": TokenType.VOLATILE,
 799        "WHEN": TokenType.WHEN,
 800        "WHERE": TokenType.WHERE,
 801        "WINDOW": TokenType.WINDOW,
 802        "WITH": TokenType.WITH,
 803        "APPLY": TokenType.APPLY,
 804        "ARRAY": TokenType.ARRAY,
 805        "BIT": TokenType.BIT,
 806        "BOOL": TokenType.BOOLEAN,
 807        "BOOLEAN": TokenType.BOOLEAN,
 808        "BYTE": TokenType.TINYINT,
 809        "MEDIUMINT": TokenType.MEDIUMINT,
 810        "INT1": TokenType.TINYINT,
 811        "TINYINT": TokenType.TINYINT,
 812        "INT16": TokenType.SMALLINT,
 813        "SHORT": TokenType.SMALLINT,
 814        "SMALLINT": TokenType.SMALLINT,
 815        "INT128": TokenType.INT128,
 816        "HUGEINT": TokenType.INT128,
 817        "INT2": TokenType.SMALLINT,
 818        "INTEGER": TokenType.INT,
 819        "INT": TokenType.INT,
 820        "INT4": TokenType.INT,
 821        "INT32": TokenType.INT,
 822        "INT64": TokenType.BIGINT,
 823        "LONG": TokenType.BIGINT,
 824        "BIGINT": TokenType.BIGINT,
 825        "INT8": TokenType.TINYINT,
 826        "UINT": TokenType.UINT,
 827        "DEC": TokenType.DECIMAL,
 828        "DECIMAL": TokenType.DECIMAL,
 829        "BIGDECIMAL": TokenType.BIGDECIMAL,
 830        "BIGNUMERIC": TokenType.BIGDECIMAL,
 831        "LIST": TokenType.LIST,
 832        "MAP": TokenType.MAP,
 833        "NULLABLE": TokenType.NULLABLE,
 834        "NUMBER": TokenType.DECIMAL,
 835        "NUMERIC": TokenType.DECIMAL,
 836        "FIXED": TokenType.DECIMAL,
 837        "REAL": TokenType.FLOAT,
 838        "FLOAT": TokenType.FLOAT,
 839        "FLOAT4": TokenType.FLOAT,
 840        "FLOAT8": TokenType.DOUBLE,
 841        "DOUBLE": TokenType.DOUBLE,
 842        "DOUBLE PRECISION": TokenType.DOUBLE,
 843        "JSON": TokenType.JSON,
 844        "JSONB": TokenType.JSONB,
 845        "CHAR": TokenType.CHAR,
 846        "CHARACTER": TokenType.CHAR,
 847        "NCHAR": TokenType.NCHAR,
 848        "VARCHAR": TokenType.VARCHAR,
 849        "VARCHAR2": TokenType.VARCHAR,
 850        "NVARCHAR": TokenType.NVARCHAR,
 851        "NVARCHAR2": TokenType.NVARCHAR,
 852        "BPCHAR": TokenType.BPCHAR,
 853        "STR": TokenType.TEXT,
 854        "STRING": TokenType.TEXT,
 855        "TEXT": TokenType.TEXT,
 856        "LONGTEXT": TokenType.LONGTEXT,
 857        "MEDIUMTEXT": TokenType.MEDIUMTEXT,
 858        "TINYTEXT": TokenType.TINYTEXT,
 859        "CLOB": TokenType.TEXT,
 860        "LONGVARCHAR": TokenType.TEXT,
 861        "BINARY": TokenType.BINARY,
 862        "BLOB": TokenType.VARBINARY,
 863        "LONGBLOB": TokenType.LONGBLOB,
 864        "MEDIUMBLOB": TokenType.MEDIUMBLOB,
 865        "TINYBLOB": TokenType.TINYBLOB,
 866        "BYTEA": TokenType.VARBINARY,
 867        "VARBINARY": TokenType.VARBINARY,
 868        "TIME": TokenType.TIME,
 869        "TIMETZ": TokenType.TIMETZ,
 870        "TIMESTAMP": TokenType.TIMESTAMP,
 871        "TIMESTAMPTZ": TokenType.TIMESTAMPTZ,
 872        "TIMESTAMPLTZ": TokenType.TIMESTAMPLTZ,
 873        "TIMESTAMP_LTZ": TokenType.TIMESTAMPLTZ,
 874        "TIMESTAMPNTZ": TokenType.TIMESTAMPNTZ,
 875        "TIMESTAMP_NTZ": TokenType.TIMESTAMPNTZ,
 876        "DATE": TokenType.DATE,
 877        "DATETIME": TokenType.DATETIME,
 878        "INT4RANGE": TokenType.INT4RANGE,
 879        "INT4MULTIRANGE": TokenType.INT4MULTIRANGE,
 880        "INT8RANGE": TokenType.INT8RANGE,
 881        "INT8MULTIRANGE": TokenType.INT8MULTIRANGE,
 882        "NUMRANGE": TokenType.NUMRANGE,
 883        "NUMMULTIRANGE": TokenType.NUMMULTIRANGE,
 884        "TSRANGE": TokenType.TSRANGE,
 885        "TSMULTIRANGE": TokenType.TSMULTIRANGE,
 886        "TSTZRANGE": TokenType.TSTZRANGE,
 887        "TSTZMULTIRANGE": TokenType.TSTZMULTIRANGE,
 888        "DATERANGE": TokenType.DATERANGE,
 889        "DATEMULTIRANGE": TokenType.DATEMULTIRANGE,
 890        "UNIQUE": TokenType.UNIQUE,
 891        "VECTOR": TokenType.VECTOR,
 892        "STRUCT": TokenType.STRUCT,
 893        "SEQUENCE": TokenType.SEQUENCE,
 894        "VARIANT": TokenType.VARIANT,
 895        "ALTER": TokenType.ALTER,
 896        "ANALYZE": TokenType.COMMAND,
 897        "CALL": TokenType.COMMAND,
 898        "COMMENT": TokenType.COMMENT,
 899        "EXPLAIN": TokenType.COMMAND,
 900        "GRANT": TokenType.COMMAND,
 901        "OPTIMIZE": TokenType.COMMAND,
 902        "PREPARE": TokenType.COMMAND,
 903        "VACUUM": TokenType.COMMAND,
 904        "USER-DEFINED": TokenType.USERDEFINED,
 905        "FOR VERSION": TokenType.VERSION_SNAPSHOT,
 906        "FOR TIMESTAMP": TokenType.TIMESTAMP_SNAPSHOT,
 907    }
 908
 909    WHITE_SPACE: t.Dict[t.Optional[str], TokenType] = {
 910        " ": TokenType.SPACE,
 911        "\t": TokenType.SPACE,
 912        "\n": TokenType.BREAK,
 913        "\r": TokenType.BREAK,
 914    }
 915
 916    COMMANDS = {
 917        TokenType.COMMAND,
 918        TokenType.EXECUTE,
 919        TokenType.FETCH,
 920        TokenType.SHOW,
 921        TokenType.RENAME,
 922    }
 923
 924    COMMAND_PREFIX_TOKENS = {TokenType.SEMICOLON, TokenType.BEGIN}
 925
 926    # Handle numeric literals like in hive (3L = BIGINT)
 927    NUMERIC_LITERALS: t.Dict[str, str] = {}
 928
 929    COMMENTS = ["--", ("/*", "*/")]
 930
 931    __slots__ = (
 932        "sql",
 933        "size",
 934        "tokens",
 935        "dialect",
 936        "_start",
 937        "_current",
 938        "_line",
 939        "_col",
 940        "_comments",
 941        "_char",
 942        "_end",
 943        "_peek",
 944        "_prev_token_line",
 945        "_rs_dialect_settings",
 946    )
 947
 948    def __init__(self, dialect: DialectType = None) -> None:
 949        from sqlglot.dialects import Dialect
 950
 951        self.dialect = Dialect.get_or_raise(dialect)
 952
 953        if USE_RS_TOKENIZER:
 954            self._rs_dialect_settings = RsTokenizerDialectSettings(
 955                unescaped_sequences=self.dialect.UNESCAPED_SEQUENCES,
 956                identifiers_can_start_with_digit=self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT,
 957            )
 958
 959        self.reset()
 960
 961    def reset(self) -> None:
 962        self.sql = ""
 963        self.size = 0
 964        self.tokens: t.List[Token] = []
 965        self._start = 0
 966        self._current = 0
 967        self._line = 1
 968        self._col = 0
 969        self._comments: t.List[str] = []
 970
 971        self._char = ""
 972        self._end = False
 973        self._peek = ""
 974        self._prev_token_line = -1
 975
 976    def tokenize(self, sql: str) -> t.List[Token]:
 977        """Returns a list of tokens corresponding to the SQL string `sql`."""
 978        if USE_RS_TOKENIZER:
 979            return self.tokenize_rs(sql)
 980
 981        self.reset()
 982        self.sql = sql
 983        self.size = len(sql)
 984
 985        try:
 986            self._scan()
 987        except Exception as e:
 988            start = max(self._current - 50, 0)
 989            end = min(self._current + 50, self.size - 1)
 990            context = self.sql[start:end]
 991            raise TokenError(f"Error tokenizing '{context}'") from e
 992
 993        return self.tokens
 994
 995    def _scan(self, until: t.Optional[t.Callable] = None) -> None:
 996        while self.size and not self._end:
 997            current = self._current
 998
 999            # Skip spaces here rather than iteratively calling advance() for performance reasons
1000            while current < self.size:
1001                char = self.sql[current]
1002
1003                if char.isspace() and (char == " " or char == "\t"):
1004                    current += 1
1005                else:
1006                    break
1007
1008            offset = current - self._current if current > self._current else 1
1009
1010            self._start = current
1011            self._advance(offset)
1012
1013            if not self._char.isspace():
1014                if self._char.isdigit():
1015                    self._scan_number()
1016                elif self._char in self._IDENTIFIERS:
1017                    self._scan_identifier(self._IDENTIFIERS[self._char])
1018                else:
1019                    self._scan_keywords()
1020
1021            if until and until():
1022                break
1023
1024        if self.tokens and self._comments:
1025            self.tokens[-1].comments.extend(self._comments)
1026
1027    def _chars(self, size: int) -> str:
1028        if size == 1:
1029            return self._char
1030
1031        start = self._current - 1
1032        end = start + size
1033
1034        return self.sql[start:end] if end <= self.size else ""
1035
1036    def _advance(self, i: int = 1, alnum: bool = False) -> None:
1037        if self.WHITE_SPACE.get(self._char) is TokenType.BREAK:
1038            # Ensures we don't count an extra line if we get a \r\n line break sequence
1039            if not (self._char == "\r" and self._peek == "\n"):
1040                self._col = 1
1041                self._line += 1
1042        else:
1043            self._col += i
1044
1045        self._current += i
1046        self._end = self._current >= self.size
1047        self._char = self.sql[self._current - 1]
1048        self._peek = "" if self._end else self.sql[self._current]
1049
1050        if alnum and self._char.isalnum():
1051            # Here we use local variables instead of attributes for better performance
1052            _col = self._col
1053            _current = self._current
1054            _end = self._end
1055            _peek = self._peek
1056
1057            while _peek.isalnum():
1058                _col += 1
1059                _current += 1
1060                _end = _current >= self.size
1061                _peek = "" if _end else self.sql[_current]
1062
1063            self._col = _col
1064            self._current = _current
1065            self._end = _end
1066            self._peek = _peek
1067            self._char = self.sql[_current - 1]
1068
1069    @property
1070    def _text(self) -> str:
1071        return self.sql[self._start : self._current]
1072
1073    def _add(self, token_type: TokenType, text: t.Optional[str] = None) -> None:
1074        self._prev_token_line = self._line
1075
1076        if self._comments and token_type == TokenType.SEMICOLON and self.tokens:
1077            self.tokens[-1].comments.extend(self._comments)
1078            self._comments = []
1079
1080        self.tokens.append(
1081            Token(
1082                token_type,
1083                text=self._text if text is None else text,
1084                line=self._line,
1085                col=self._col,
1086                start=self._start,
1087                end=self._current - 1,
1088                comments=self._comments,
1089            )
1090        )
1091        self._comments = []
1092
1093        # If we have either a semicolon or a begin token before the command's token, we'll parse
1094        # whatever follows the command's token as a string
1095        if (
1096            token_type in self.COMMANDS
1097            and self._peek != ";"
1098            and (len(self.tokens) == 1 or self.tokens[-2].token_type in self.COMMAND_PREFIX_TOKENS)
1099        ):
1100            start = self._current
1101            tokens = len(self.tokens)
1102            self._scan(lambda: self._peek == ";")
1103            self.tokens = self.tokens[:tokens]
1104            text = self.sql[start : self._current].strip()
1105            if text:
1106                self._add(TokenType.STRING, text)
1107
1108    def _scan_keywords(self) -> None:
1109        size = 0
1110        word = None
1111        chars = self._text
1112        char = chars
1113        prev_space = False
1114        skip = False
1115        trie = self._KEYWORD_TRIE
1116        single_token = char in self.SINGLE_TOKENS
1117
1118        while chars:
1119            if skip:
1120                result = TrieResult.PREFIX
1121            else:
1122                result, trie = in_trie(trie, char.upper())
1123
1124            if result == TrieResult.FAILED:
1125                break
1126            if result == TrieResult.EXISTS:
1127                word = chars
1128
1129            end = self._current + size
1130            size += 1
1131
1132            if end < self.size:
1133                char = self.sql[end]
1134                single_token = single_token or char in self.SINGLE_TOKENS
1135                is_space = char.isspace()
1136
1137                if not is_space or not prev_space:
1138                    if is_space:
1139                        char = " "
1140                    chars += char
1141                    prev_space = is_space
1142                    skip = False
1143                else:
1144                    skip = True
1145            else:
1146                char = ""
1147                break
1148
1149        if word:
1150            if self._scan_string(word):
1151                return
1152            if self._scan_comment(word):
1153                return
1154            if prev_space or single_token or not char:
1155                self._advance(size - 1)
1156                word = word.upper()
1157                self._add(self.KEYWORDS[word], text=word)
1158                return
1159
1160        if self._char in self.SINGLE_TOKENS:
1161            self._add(self.SINGLE_TOKENS[self._char], text=self._char)
1162            return
1163
1164        self._scan_var()
1165
1166    def _scan_comment(self, comment_start: str) -> bool:
1167        if comment_start not in self._COMMENTS:
1168            return False
1169
1170        comment_start_line = self._line
1171        comment_start_size = len(comment_start)
1172        comment_end = self._COMMENTS[comment_start]
1173
1174        if comment_end:
1175            # Skip the comment's start delimiter
1176            self._advance(comment_start_size)
1177
1178            comment_count = 1
1179            comment_end_size = len(comment_end)
1180
1181            while not self._end:
1182                if self._chars(comment_end_size) == comment_end:
1183                    comment_count -= 1
1184                    if not comment_count:
1185                        break
1186
1187                self._advance(alnum=True)
1188
1189                # Nested comments are allowed by some dialects, e.g. databricks, duckdb, postgres
1190                if (
1191                    self.NESTED_COMMENTS
1192                    and not self._end
1193                    and self._chars(comment_end_size) == comment_start
1194                ):
1195                    self._advance(comment_start_size)
1196                    comment_count += 1
1197
1198            self._comments.append(self._text[comment_start_size : -comment_end_size + 1])
1199            self._advance(comment_end_size - 1)
1200        else:
1201            while not self._end and self.WHITE_SPACE.get(self._peek) is not TokenType.BREAK:
1202                self._advance(alnum=True)
1203            self._comments.append(self._text[comment_start_size:])
1204
1205        # Leading comment is attached to the succeeding token, whilst trailing comment to the preceding.
1206        # Multiple consecutive comments are preserved by appending them to the current comments list.
1207        if comment_start_line == self._prev_token_line:
1208            self.tokens[-1].comments.extend(self._comments)
1209            self._comments = []
1210            self._prev_token_line = self._line
1211
1212        return True
1213
1214    def _scan_number(self) -> None:
1215        if self._char == "0":
1216            peek = self._peek.upper()
1217            if peek == "B":
1218                return self._scan_bits() if self.BIT_STRINGS else self._add(TokenType.NUMBER)
1219            elif peek == "X":
1220                return self._scan_hex() if self.HEX_STRINGS else self._add(TokenType.NUMBER)
1221
1222        decimal = False
1223        scientific = 0
1224
1225        while True:
1226            if self._peek.isdigit():
1227                self._advance()
1228            elif self._peek == "." and not decimal:
1229                if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER:
1230                    return self._add(TokenType.NUMBER)
1231                decimal = True
1232                self._advance()
1233            elif self._peek in ("-", "+") and scientific == 1:
1234                scientific += 1
1235                self._advance()
1236            elif self._peek.upper() == "E" and not scientific:
1237                scientific += 1
1238                self._advance()
1239            elif self._peek.isidentifier():
1240                number_text = self._text
1241                literal = ""
1242
1243                while self._peek.strip() and self._peek not in self.SINGLE_TOKENS:
1244                    literal += self._peek
1245                    self._advance()
1246
1247                token_type = self.KEYWORDS.get(self.NUMERIC_LITERALS.get(literal.upper(), ""))
1248
1249                if token_type:
1250                    self._add(TokenType.NUMBER, number_text)
1251                    self._add(TokenType.DCOLON, "::")
1252                    return self._add(token_type, literal)
1253                elif self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT:
1254                    return self._add(TokenType.VAR)
1255
1256                self._advance(-len(literal))
1257                return self._add(TokenType.NUMBER, number_text)
1258            else:
1259                return self._add(TokenType.NUMBER)
1260
1261    def _scan_bits(self) -> None:
1262        self._advance()
1263        value = self._extract_value()
1264        try:
1265            # If `value` can't be converted to a binary, fallback to tokenizing it as an identifier
1266            int(value, 2)
1267            self._add(TokenType.BIT_STRING, value[2:])  # Drop the 0b
1268        except ValueError:
1269            self._add(TokenType.IDENTIFIER)
1270
1271    def _scan_hex(self) -> None:
1272        self._advance()
1273        value = self._extract_value()
1274        try:
1275            # If `value` can't be converted to a hex, fallback to tokenizing it as an identifier
1276            int(value, 16)
1277            self._add(TokenType.HEX_STRING, value[2:])  # Drop the 0x
1278        except ValueError:
1279            self._add(TokenType.IDENTIFIER)
1280
1281    def _extract_value(self) -> str:
1282        while True:
1283            char = self._peek.strip()
1284            if char and char not in self.SINGLE_TOKENS:
1285                self._advance(alnum=True)
1286            else:
1287                break
1288
1289        return self._text
1290
1291    def _scan_string(self, start: str) -> bool:
1292        base = None
1293        token_type = TokenType.STRING
1294
1295        if start in self._QUOTES:
1296            end = self._QUOTES[start]
1297        elif start in self._FORMAT_STRINGS:
1298            end, token_type = self._FORMAT_STRINGS[start]
1299
1300            if token_type == TokenType.HEX_STRING:
1301                base = 16
1302            elif token_type == TokenType.BIT_STRING:
1303                base = 2
1304            elif token_type == TokenType.HEREDOC_STRING:
1305                self._advance()
1306
1307                if self._char == end:
1308                    tag = ""
1309                else:
1310                    tag = self._extract_string(
1311                        end,
1312                        raw_string=True,
1313                        raise_unmatched=not self.HEREDOC_TAG_IS_IDENTIFIER,
1314                    )
1315
1316                if tag and self.HEREDOC_TAG_IS_IDENTIFIER and (self._end or not tag.isidentifier()):
1317                    if not self._end:
1318                        self._advance(-1)
1319
1320                    self._advance(-len(tag))
1321                    self._add(self.HEREDOC_STRING_ALTERNATIVE)
1322                    return True
1323
1324                end = f"{start}{tag}{end}"
1325        else:
1326            return False
1327
1328        self._advance(len(start))
1329        text = self._extract_string(end, raw_string=token_type == TokenType.RAW_STRING)
1330
1331        if base:
1332            try:
1333                int(text, base)
1334            except Exception:
1335                raise TokenError(
1336                    f"Numeric string contains invalid characters from {self._line}:{self._start}"
1337                )
1338
1339        self._add(token_type, text)
1340        return True
1341
1342    def _scan_identifier(self, identifier_end: str) -> None:
1343        self._advance()
1344        text = self._extract_string(identifier_end, escapes=self._IDENTIFIER_ESCAPES)
1345        self._add(TokenType.IDENTIFIER, text)
1346
1347    def _scan_var(self) -> None:
1348        while True:
1349            char = self._peek.strip()
1350            if char and (char in self.VAR_SINGLE_TOKENS or char not in self.SINGLE_TOKENS):
1351                self._advance(alnum=True)
1352            else:
1353                break
1354
1355        self._add(
1356            TokenType.VAR
1357            if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER
1358            else self.KEYWORDS.get(self._text.upper(), TokenType.VAR)
1359        )
1360
1361    def _extract_string(
1362        self,
1363        delimiter: str,
1364        escapes: t.Optional[t.Set[str]] = None,
1365        raw_string: bool = False,
1366        raise_unmatched: bool = True,
1367    ) -> str:
1368        text = ""
1369        delim_size = len(delimiter)
1370        escapes = self._STRING_ESCAPES if escapes is None else escapes
1371
1372        while True:
1373            if (
1374                not raw_string
1375                and self.dialect.UNESCAPED_SEQUENCES
1376                and self._peek
1377                and self._char in self.STRING_ESCAPES
1378            ):
1379                unescaped_sequence = self.dialect.UNESCAPED_SEQUENCES.get(self._char + self._peek)
1380                if unescaped_sequence:
1381                    self._advance(2)
1382                    text += unescaped_sequence
1383                    continue
1384            if (
1385                (self.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS or not raw_string)
1386                and self._char in escapes
1387                and (self._peek == delimiter or self._peek in escapes)
1388                and (self._char not in self._QUOTES or self._char == self._peek)
1389            ):
1390                if self._peek == delimiter:
1391                    text += self._peek
1392                else:
1393                    text += self._char + self._peek
1394
1395                if self._current + 1 < self.size:
1396                    self._advance(2)
1397                else:
1398                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._current}")
1399            else:
1400                if self._chars(delim_size) == delimiter:
1401                    if delim_size > 1:
1402                        self._advance(delim_size - 1)
1403                    break
1404
1405                if self._end:
1406                    if not raise_unmatched:
1407                        return text + self._char
1408
1409                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._start}")
1410
1411                current = self._current - 1
1412                self._advance(alnum=True)
1413                text += self.sql[current : self._current - 1]
1414
1415        return text
1416
1417    def tokenize_rs(self, sql: str) -> t.List[Token]:
1418        if not self._RS_TOKENIZER:
1419            raise SqlglotError("Rust tokenizer is not available")
1420
1421        try:
1422            tokens = self._RS_TOKENIZER.tokenize(sql, self._rs_dialect_settings)
1423            for token in tokens:
1424                token.token_type = _ALL_TOKEN_TYPES[token.token_type_index]
1425            return tokens
1426        except Exception as e:
1427            raise TokenError(str(e))