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    DECIMAL32 = auto()
 127    DECIMAL64 = auto()
 128    DECIMAL128 = auto()
 129    UDECIMAL = auto()
 130    BIGDECIMAL = auto()
 131    CHAR = auto()
 132    NCHAR = auto()
 133    VARCHAR = auto()
 134    NVARCHAR = auto()
 135    BPCHAR = auto()
 136    TEXT = auto()
 137    MEDIUMTEXT = auto()
 138    LONGTEXT = auto()
 139    MEDIUMBLOB = auto()
 140    LONGBLOB = auto()
 141    TINYBLOB = auto()
 142    TINYTEXT = auto()
 143    NAME = auto()
 144    BINARY = auto()
 145    VARBINARY = auto()
 146    JSON = auto()
 147    JSONB = auto()
 148    TIME = auto()
 149    TIMETZ = auto()
 150    TIMESTAMP = auto()
 151    TIMESTAMPTZ = auto()
 152    TIMESTAMPLTZ = auto()
 153    TIMESTAMPNTZ = auto()
 154    TIMESTAMP_S = auto()
 155    TIMESTAMP_MS = auto()
 156    TIMESTAMP_NS = auto()
 157    DATETIME = auto()
 158    DATETIME64 = auto()
 159    DATE = auto()
 160    DATE32 = auto()
 161    INT4RANGE = auto()
 162    INT4MULTIRANGE = auto()
 163    INT8RANGE = auto()
 164    INT8MULTIRANGE = auto()
 165    NUMRANGE = auto()
 166    NUMMULTIRANGE = auto()
 167    TSRANGE = auto()
 168    TSMULTIRANGE = auto()
 169    TSTZRANGE = auto()
 170    TSTZMULTIRANGE = auto()
 171    DATERANGE = auto()
 172    DATEMULTIRANGE = auto()
 173    UUID = auto()
 174    GEOGRAPHY = auto()
 175    NULLABLE = auto()
 176    GEOMETRY = auto()
 177    HLLSKETCH = auto()
 178    HSTORE = auto()
 179    SUPER = auto()
 180    SERIAL = auto()
 181    SMALLSERIAL = auto()
 182    BIGSERIAL = auto()
 183    XML = auto()
 184    YEAR = auto()
 185    UNIQUEIDENTIFIER = auto()
 186    USERDEFINED = auto()
 187    MONEY = auto()
 188    SMALLMONEY = auto()
 189    ROWVERSION = auto()
 190    IMAGE = auto()
 191    VARIANT = auto()
 192    OBJECT = auto()
 193    INET = auto()
 194    IPADDRESS = auto()
 195    IPPREFIX = auto()
 196    IPV4 = auto()
 197    IPV6 = auto()
 198    ENUM = auto()
 199    ENUM8 = auto()
 200    ENUM16 = auto()
 201    FIXEDSTRING = auto()
 202    LOWCARDINALITY = auto()
 203    NESTED = auto()
 204    AGGREGATEFUNCTION = auto()
 205    SIMPLEAGGREGATEFUNCTION = auto()
 206    TDIGEST = auto()
 207    UNKNOWN = auto()
 208    VECTOR = auto()
 209
 210    # keywords
 211    ALIAS = auto()
 212    ALTER = auto()
 213    ALWAYS = auto()
 214    ALL = auto()
 215    ANTI = auto()
 216    ANY = auto()
 217    APPLY = auto()
 218    ARRAY = auto()
 219    ASC = auto()
 220    ASOF = auto()
 221    AUTO_INCREMENT = auto()
 222    BEGIN = auto()
 223    BETWEEN = auto()
 224    CACHE = auto()
 225    CASE = auto()
 226    CHARACTER_SET = auto()
 227    CLUSTER_BY = auto()
 228    COLLATE = auto()
 229    COMMAND = auto()
 230    COMMENT = auto()
 231    COMMIT = auto()
 232    CONNECT_BY = auto()
 233    CONSTRAINT = auto()
 234    COPY = auto()
 235    CREATE = auto()
 236    CROSS = auto()
 237    CUBE = auto()
 238    CURRENT_DATE = auto()
 239    CURRENT_DATETIME = auto()
 240    CURRENT_TIME = auto()
 241    CURRENT_TIMESTAMP = auto()
 242    CURRENT_USER = auto()
 243    DECLARE = auto()
 244    DEFAULT = auto()
 245    DELETE = auto()
 246    DESC = auto()
 247    DESCRIBE = auto()
 248    DICTIONARY = auto()
 249    DISTINCT = auto()
 250    DISTRIBUTE_BY = auto()
 251    DIV = auto()
 252    DROP = auto()
 253    ELSE = auto()
 254    END = auto()
 255    ESCAPE = auto()
 256    EXCEPT = auto()
 257    EXECUTE = auto()
 258    EXISTS = auto()
 259    FALSE = auto()
 260    FETCH = auto()
 261    FILTER = auto()
 262    FINAL = auto()
 263    FIRST = auto()
 264    FOR = auto()
 265    FORCE = auto()
 266    FOREIGN_KEY = auto()
 267    FORMAT = auto()
 268    FROM = auto()
 269    FULL = auto()
 270    FUNCTION = auto()
 271    GLOB = auto()
 272    GLOBAL = auto()
 273    GRANT = auto()
 274    GROUP_BY = auto()
 275    GROUPING_SETS = auto()
 276    HAVING = auto()
 277    HINT = auto()
 278    IGNORE = auto()
 279    ILIKE = auto()
 280    ILIKE_ANY = auto()
 281    IN = auto()
 282    INDEX = auto()
 283    INNER = auto()
 284    INSERT = auto()
 285    INTERSECT = auto()
 286    INTERVAL = auto()
 287    INTO = auto()
 288    INTRODUCER = auto()
 289    IRLIKE = auto()
 290    IS = auto()
 291    ISNULL = auto()
 292    JOIN = auto()
 293    JOIN_MARKER = auto()
 294    KEEP = auto()
 295    KEY = auto()
 296    KILL = auto()
 297    LANGUAGE = auto()
 298    LATERAL = auto()
 299    LEFT = auto()
 300    LIKE = auto()
 301    LIKE_ANY = auto()
 302    LIMIT = auto()
 303    LIST = auto()
 304    LOAD = auto()
 305    LOCK = auto()
 306    MAP = auto()
 307    MATCH_CONDITION = auto()
 308    MATCH_RECOGNIZE = auto()
 309    MEMBER_OF = auto()
 310    MERGE = auto()
 311    MOD = auto()
 312    MODEL = auto()
 313    NATURAL = auto()
 314    NEXT = auto()
 315    NOTNULL = auto()
 316    NULL = auto()
 317    OBJECT_IDENTIFIER = auto()
 318    OFFSET = auto()
 319    ON = auto()
 320    ONLY = auto()
 321    OPERATOR = auto()
 322    ORDER_BY = auto()
 323    ORDER_SIBLINGS_BY = auto()
 324    ORDERED = auto()
 325    ORDINALITY = auto()
 326    OUTER = auto()
 327    OVER = auto()
 328    OVERLAPS = auto()
 329    OVERWRITE = auto()
 330    PARTITION = auto()
 331    PARTITION_BY = auto()
 332    PERCENT = auto()
 333    PIVOT = auto()
 334    PLACEHOLDER = auto()
 335    POSITIONAL = auto()
 336    PRAGMA = auto()
 337    PREWHERE = auto()
 338    PRIMARY_KEY = auto()
 339    PROCEDURE = auto()
 340    PROPERTIES = auto()
 341    PSEUDO_TYPE = auto()
 342    QUALIFY = auto()
 343    QUOTE = auto()
 344    RANGE = auto()
 345    RECURSIVE = auto()
 346    REFRESH = auto()
 347    RENAME = auto()
 348    REPLACE = auto()
 349    RETURNING = auto()
 350    REFERENCES = auto()
 351    RIGHT = auto()
 352    RLIKE = auto()
 353    ROLLBACK = auto()
 354    ROLLUP = auto()
 355    ROW = auto()
 356    ROWS = auto()
 357    SELECT = auto()
 358    SEMI = auto()
 359    SEPARATOR = auto()
 360    SEQUENCE = auto()
 361    SERDE_PROPERTIES = auto()
 362    SET = auto()
 363    SETTINGS = auto()
 364    SHOW = auto()
 365    SIMILAR_TO = auto()
 366    SOME = auto()
 367    SORT_BY = auto()
 368    START_WITH = auto()
 369    STORAGE_INTEGRATION = auto()
 370    STRAIGHT_JOIN = auto()
 371    STRUCT = auto()
 372    SUMMARIZE = auto()
 373    TABLE_SAMPLE = auto()
 374    TAG = auto()
 375    TEMPORARY = auto()
 376    TOP = auto()
 377    THEN = auto()
 378    TRUE = auto()
 379    TRUNCATE = auto()
 380    UNCACHE = auto()
 381    UNION = auto()
 382    UNNEST = auto()
 383    UNPIVOT = auto()
 384    UPDATE = auto()
 385    USE = auto()
 386    USING = auto()
 387    VALUES = auto()
 388    VIEW = auto()
 389    VOLATILE = auto()
 390    WHEN = auto()
 391    WHERE = auto()
 392    WINDOW = auto()
 393    WITH = auto()
 394    UNIQUE = auto()
 395    VERSION_SNAPSHOT = auto()
 396    TIMESTAMP_SNAPSHOT = auto()
 397    OPTION = auto()
 398
 399
 400_ALL_TOKEN_TYPES = list(TokenType)
 401_TOKEN_TYPE_TO_INDEX = {token_type: i for i, token_type in enumerate(_ALL_TOKEN_TYPES)}
 402
 403
 404class Token:
 405    __slots__ = ("token_type", "text", "line", "col", "start", "end", "comments")
 406
 407    @classmethod
 408    def number(cls, number: int) -> Token:
 409        """Returns a NUMBER token with `number` as its text."""
 410        return cls(TokenType.NUMBER, str(number))
 411
 412    @classmethod
 413    def string(cls, string: str) -> Token:
 414        """Returns a STRING token with `string` as its text."""
 415        return cls(TokenType.STRING, string)
 416
 417    @classmethod
 418    def identifier(cls, identifier: str) -> Token:
 419        """Returns an IDENTIFIER token with `identifier` as its text."""
 420        return cls(TokenType.IDENTIFIER, identifier)
 421
 422    @classmethod
 423    def var(cls, var: str) -> Token:
 424        """Returns an VAR token with `var` as its text."""
 425        return cls(TokenType.VAR, var)
 426
 427    def __init__(
 428        self,
 429        token_type: TokenType,
 430        text: str,
 431        line: int = 1,
 432        col: int = 1,
 433        start: int = 0,
 434        end: int = 0,
 435        comments: t.Optional[t.List[str]] = None,
 436    ) -> None:
 437        """Token initializer.
 438
 439        Args:
 440            token_type: The TokenType Enum.
 441            text: The text of the token.
 442            line: The line that the token ends on.
 443            col: The column that the token ends on.
 444            start: The start index of the token.
 445            end: The ending index of the token.
 446            comments: The comments to attach to the token.
 447        """
 448        self.token_type = token_type
 449        self.text = text
 450        self.line = line
 451        self.col = col
 452        self.start = start
 453        self.end = end
 454        self.comments = [] if comments is None else comments
 455
 456    def __repr__(self) -> str:
 457        attributes = ", ".join(f"{k}: {getattr(self, k)}" for k in self.__slots__)
 458        return f"<Token {attributes}>"
 459
 460
 461class _Tokenizer(type):
 462    def __new__(cls, clsname, bases, attrs):
 463        klass = super().__new__(cls, clsname, bases, attrs)
 464
 465        def _convert_quotes(arr: t.List[str | t.Tuple[str, str]]) -> t.Dict[str, str]:
 466            return dict(
 467                (item, item) if isinstance(item, str) else (item[0], item[1]) for item in arr
 468            )
 469
 470        def _quotes_to_format(
 471            token_type: TokenType, arr: t.List[str | t.Tuple[str, str]]
 472        ) -> t.Dict[str, t.Tuple[str, TokenType]]:
 473            return {k: (v, token_type) for k, v in _convert_quotes(arr).items()}
 474
 475        klass._QUOTES = _convert_quotes(klass.QUOTES)
 476        klass._IDENTIFIERS = _convert_quotes(klass.IDENTIFIERS)
 477
 478        klass._FORMAT_STRINGS = {
 479            **{
 480                p + s: (e, TokenType.NATIONAL_STRING)
 481                for s, e in klass._QUOTES.items()
 482                for p in ("n", "N")
 483            },
 484            **_quotes_to_format(TokenType.BIT_STRING, klass.BIT_STRINGS),
 485            **_quotes_to_format(TokenType.BYTE_STRING, klass.BYTE_STRINGS),
 486            **_quotes_to_format(TokenType.HEX_STRING, klass.HEX_STRINGS),
 487            **_quotes_to_format(TokenType.RAW_STRING, klass.RAW_STRINGS),
 488            **_quotes_to_format(TokenType.HEREDOC_STRING, klass.HEREDOC_STRINGS),
 489            **_quotes_to_format(TokenType.UNICODE_STRING, klass.UNICODE_STRINGS),
 490        }
 491
 492        klass._STRING_ESCAPES = set(klass.STRING_ESCAPES)
 493        klass._IDENTIFIER_ESCAPES = set(klass.IDENTIFIER_ESCAPES)
 494        klass._COMMENTS = {
 495            **dict(
 496                (comment, None) if isinstance(comment, str) else (comment[0], comment[1])
 497                for comment in klass.COMMENTS
 498            ),
 499            "{#": "#}",  # Ensure Jinja comments are tokenized correctly in all dialects
 500        }
 501
 502        klass._KEYWORD_TRIE = new_trie(
 503            key.upper()
 504            for key in (
 505                *klass.KEYWORDS,
 506                *klass._COMMENTS,
 507                *klass._QUOTES,
 508                *klass._FORMAT_STRINGS,
 509            )
 510            if " " in key or any(single in key for single in klass.SINGLE_TOKENS)
 511        )
 512
 513        if USE_RS_TOKENIZER:
 514            settings = RsTokenizerSettings(
 515                white_space={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.WHITE_SPACE.items()},
 516                single_tokens={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.SINGLE_TOKENS.items()},
 517                keywords={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.KEYWORDS.items()},
 518                numeric_literals=klass.NUMERIC_LITERALS,
 519                identifiers=klass._IDENTIFIERS,
 520                identifier_escapes=klass._IDENTIFIER_ESCAPES,
 521                string_escapes=klass._STRING_ESCAPES,
 522                quotes=klass._QUOTES,
 523                format_strings={
 524                    k: (v1, _TOKEN_TYPE_TO_INDEX[v2])
 525                    for k, (v1, v2) in klass._FORMAT_STRINGS.items()
 526                },
 527                has_bit_strings=bool(klass.BIT_STRINGS),
 528                has_hex_strings=bool(klass.HEX_STRINGS),
 529                comments=klass._COMMENTS,
 530                var_single_tokens=klass.VAR_SINGLE_TOKENS,
 531                commands={_TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMANDS},
 532                command_prefix_tokens={
 533                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMAND_PREFIX_TOKENS
 534                },
 535                heredoc_tag_is_identifier=klass.HEREDOC_TAG_IS_IDENTIFIER,
 536                string_escapes_allowed_in_raw_strings=klass.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS,
 537                nested_comments=klass.NESTED_COMMENTS,
 538            )
 539            token_types = RsTokenTypeSettings(
 540                bit_string=_TOKEN_TYPE_TO_INDEX[TokenType.BIT_STRING],
 541                break_=_TOKEN_TYPE_TO_INDEX[TokenType.BREAK],
 542                dcolon=_TOKEN_TYPE_TO_INDEX[TokenType.DCOLON],
 543                heredoc_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEREDOC_STRING],
 544                raw_string=_TOKEN_TYPE_TO_INDEX[TokenType.RAW_STRING],
 545                hex_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEX_STRING],
 546                identifier=_TOKEN_TYPE_TO_INDEX[TokenType.IDENTIFIER],
 547                number=_TOKEN_TYPE_TO_INDEX[TokenType.NUMBER],
 548                parameter=_TOKEN_TYPE_TO_INDEX[TokenType.PARAMETER],
 549                semicolon=_TOKEN_TYPE_TO_INDEX[TokenType.SEMICOLON],
 550                string=_TOKEN_TYPE_TO_INDEX[TokenType.STRING],
 551                var=_TOKEN_TYPE_TO_INDEX[TokenType.VAR],
 552                heredoc_string_alternative=_TOKEN_TYPE_TO_INDEX[klass.HEREDOC_STRING_ALTERNATIVE],
 553            )
 554            klass._RS_TOKENIZER = RsTokenizer(settings, token_types)
 555        else:
 556            klass._RS_TOKENIZER = None
 557
 558        return klass
 559
 560
 561class Tokenizer(metaclass=_Tokenizer):
 562    SINGLE_TOKENS = {
 563        "(": TokenType.L_PAREN,
 564        ")": TokenType.R_PAREN,
 565        "[": TokenType.L_BRACKET,
 566        "]": TokenType.R_BRACKET,
 567        "{": TokenType.L_BRACE,
 568        "}": TokenType.R_BRACE,
 569        "&": TokenType.AMP,
 570        "^": TokenType.CARET,
 571        ":": TokenType.COLON,
 572        ",": TokenType.COMMA,
 573        ".": TokenType.DOT,
 574        "-": TokenType.DASH,
 575        "=": TokenType.EQ,
 576        ">": TokenType.GT,
 577        "<": TokenType.LT,
 578        "%": TokenType.MOD,
 579        "!": TokenType.NOT,
 580        "|": TokenType.PIPE,
 581        "+": TokenType.PLUS,
 582        ";": TokenType.SEMICOLON,
 583        "/": TokenType.SLASH,
 584        "\\": TokenType.BACKSLASH,
 585        "*": TokenType.STAR,
 586        "~": TokenType.TILDA,
 587        "?": TokenType.PLACEHOLDER,
 588        "@": TokenType.PARAMETER,
 589        "#": TokenType.HASH,
 590        # Used for breaking a var like x'y' but nothing else the token type doesn't matter
 591        "'": TokenType.UNKNOWN,
 592        "`": TokenType.UNKNOWN,
 593        '"': TokenType.UNKNOWN,
 594    }
 595
 596    BIT_STRINGS: t.List[str | t.Tuple[str, str]] = []
 597    BYTE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 598    HEX_STRINGS: t.List[str | t.Tuple[str, str]] = []
 599    RAW_STRINGS: t.List[str | t.Tuple[str, str]] = []
 600    HEREDOC_STRINGS: t.List[str | t.Tuple[str, str]] = []
 601    UNICODE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 602    IDENTIFIERS: t.List[str | t.Tuple[str, str]] = ['"']
 603    IDENTIFIER_ESCAPES = ['"']
 604    QUOTES: t.List[t.Tuple[str, str] | str] = ["'"]
 605    STRING_ESCAPES = ["'"]
 606    VAR_SINGLE_TOKENS: t.Set[str] = set()
 607
 608    # Whether the heredoc tags follow the same lexical rules as unquoted identifiers
 609    HEREDOC_TAG_IS_IDENTIFIER = False
 610
 611    # Token that we'll generate as a fallback if the heredoc prefix doesn't correspond to a heredoc
 612    HEREDOC_STRING_ALTERNATIVE = TokenType.VAR
 613
 614    # Whether string escape characters function as such when placed within raw strings
 615    STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS = True
 616
 617    NESTED_COMMENTS = True
 618
 619    # Autofilled
 620    _COMMENTS: t.Dict[str, str] = {}
 621    _FORMAT_STRINGS: t.Dict[str, t.Tuple[str, TokenType]] = {}
 622    _IDENTIFIERS: t.Dict[str, str] = {}
 623    _IDENTIFIER_ESCAPES: t.Set[str] = set()
 624    _QUOTES: t.Dict[str, str] = {}
 625    _STRING_ESCAPES: t.Set[str] = set()
 626    _KEYWORD_TRIE: t.Dict = {}
 627    _RS_TOKENIZER: t.Optional[t.Any] = None
 628
 629    KEYWORDS: t.Dict[str, TokenType] = {
 630        **{f"{{%{postfix}": TokenType.BLOCK_START for postfix in ("", "+", "-")},
 631        **{f"{prefix}%}}": TokenType.BLOCK_END for prefix in ("", "+", "-")},
 632        **{f"{{{{{postfix}": TokenType.BLOCK_START for postfix in ("+", "-")},
 633        **{f"{prefix}}}}}": TokenType.BLOCK_END for prefix in ("+", "-")},
 634        "/*+": TokenType.HINT,
 635        "==": TokenType.EQ,
 636        "::": TokenType.DCOLON,
 637        "||": TokenType.DPIPE,
 638        ">=": TokenType.GTE,
 639        "<=": TokenType.LTE,
 640        "<>": TokenType.NEQ,
 641        "!=": TokenType.NEQ,
 642        ":=": TokenType.COLON_EQ,
 643        "<=>": TokenType.NULLSAFE_EQ,
 644        "->": TokenType.ARROW,
 645        "->>": TokenType.DARROW,
 646        "=>": TokenType.FARROW,
 647        "#>": TokenType.HASH_ARROW,
 648        "#>>": TokenType.DHASH_ARROW,
 649        "<->": TokenType.LR_ARROW,
 650        "&&": TokenType.DAMP,
 651        "??": TokenType.DQMARK,
 652        "ALL": TokenType.ALL,
 653        "ALWAYS": TokenType.ALWAYS,
 654        "AND": TokenType.AND,
 655        "ANTI": TokenType.ANTI,
 656        "ANY": TokenType.ANY,
 657        "ASC": TokenType.ASC,
 658        "AS": TokenType.ALIAS,
 659        "ASOF": TokenType.ASOF,
 660        "AUTOINCREMENT": TokenType.AUTO_INCREMENT,
 661        "AUTO_INCREMENT": TokenType.AUTO_INCREMENT,
 662        "BEGIN": TokenType.BEGIN,
 663        "BETWEEN": TokenType.BETWEEN,
 664        "CACHE": TokenType.CACHE,
 665        "UNCACHE": TokenType.UNCACHE,
 666        "CASE": TokenType.CASE,
 667        "CHARACTER SET": TokenType.CHARACTER_SET,
 668        "CLUSTER BY": TokenType.CLUSTER_BY,
 669        "COLLATE": TokenType.COLLATE,
 670        "COLUMN": TokenType.COLUMN,
 671        "COMMIT": TokenType.COMMIT,
 672        "CONNECT BY": TokenType.CONNECT_BY,
 673        "CONSTRAINT": TokenType.CONSTRAINT,
 674        "COPY": TokenType.COPY,
 675        "CREATE": TokenType.CREATE,
 676        "CROSS": TokenType.CROSS,
 677        "CUBE": TokenType.CUBE,
 678        "CURRENT_DATE": TokenType.CURRENT_DATE,
 679        "CURRENT_TIME": TokenType.CURRENT_TIME,
 680        "CURRENT_TIMESTAMP": TokenType.CURRENT_TIMESTAMP,
 681        "CURRENT_USER": TokenType.CURRENT_USER,
 682        "DATABASE": TokenType.DATABASE,
 683        "DEFAULT": TokenType.DEFAULT,
 684        "DELETE": TokenType.DELETE,
 685        "DESC": TokenType.DESC,
 686        "DESCRIBE": TokenType.DESCRIBE,
 687        "DISTINCT": TokenType.DISTINCT,
 688        "DISTRIBUTE BY": TokenType.DISTRIBUTE_BY,
 689        "DIV": TokenType.DIV,
 690        "DROP": TokenType.DROP,
 691        "ELSE": TokenType.ELSE,
 692        "END": TokenType.END,
 693        "ENUM": TokenType.ENUM,
 694        "ESCAPE": TokenType.ESCAPE,
 695        "EXCEPT": TokenType.EXCEPT,
 696        "EXECUTE": TokenType.EXECUTE,
 697        "EXISTS": TokenType.EXISTS,
 698        "FALSE": TokenType.FALSE,
 699        "FETCH": TokenType.FETCH,
 700        "FILTER": TokenType.FILTER,
 701        "FIRST": TokenType.FIRST,
 702        "FULL": TokenType.FULL,
 703        "FUNCTION": TokenType.FUNCTION,
 704        "FOR": TokenType.FOR,
 705        "FOREIGN KEY": TokenType.FOREIGN_KEY,
 706        "FORMAT": TokenType.FORMAT,
 707        "FROM": TokenType.FROM,
 708        "GEOGRAPHY": TokenType.GEOGRAPHY,
 709        "GEOMETRY": TokenType.GEOMETRY,
 710        "GLOB": TokenType.GLOB,
 711        "GROUP BY": TokenType.GROUP_BY,
 712        "GROUPING SETS": TokenType.GROUPING_SETS,
 713        "HAVING": TokenType.HAVING,
 714        "ILIKE": TokenType.ILIKE,
 715        "IN": TokenType.IN,
 716        "INDEX": TokenType.INDEX,
 717        "INET": TokenType.INET,
 718        "INNER": TokenType.INNER,
 719        "INSERT": TokenType.INSERT,
 720        "INTERVAL": TokenType.INTERVAL,
 721        "INTERSECT": TokenType.INTERSECT,
 722        "INTO": TokenType.INTO,
 723        "IS": TokenType.IS,
 724        "ISNULL": TokenType.ISNULL,
 725        "JOIN": TokenType.JOIN,
 726        "KEEP": TokenType.KEEP,
 727        "KILL": TokenType.KILL,
 728        "LATERAL": TokenType.LATERAL,
 729        "LEFT": TokenType.LEFT,
 730        "LIKE": TokenType.LIKE,
 731        "LIMIT": TokenType.LIMIT,
 732        "LOAD": TokenType.LOAD,
 733        "LOCK": TokenType.LOCK,
 734        "MERGE": TokenType.MERGE,
 735        "NATURAL": TokenType.NATURAL,
 736        "NEXT": TokenType.NEXT,
 737        "NOT": TokenType.NOT,
 738        "NOTNULL": TokenType.NOTNULL,
 739        "NULL": TokenType.NULL,
 740        "OBJECT": TokenType.OBJECT,
 741        "OFFSET": TokenType.OFFSET,
 742        "ON": TokenType.ON,
 743        "OR": TokenType.OR,
 744        "XOR": TokenType.XOR,
 745        "ORDER BY": TokenType.ORDER_BY,
 746        "ORDINALITY": TokenType.ORDINALITY,
 747        "OUTER": TokenType.OUTER,
 748        "OVER": TokenType.OVER,
 749        "OVERLAPS": TokenType.OVERLAPS,
 750        "OVERWRITE": TokenType.OVERWRITE,
 751        "PARTITION": TokenType.PARTITION,
 752        "PARTITION BY": TokenType.PARTITION_BY,
 753        "PARTITIONED BY": TokenType.PARTITION_BY,
 754        "PARTITIONED_BY": TokenType.PARTITION_BY,
 755        "PERCENT": TokenType.PERCENT,
 756        "PIVOT": TokenType.PIVOT,
 757        "PRAGMA": TokenType.PRAGMA,
 758        "PRIMARY KEY": TokenType.PRIMARY_KEY,
 759        "PROCEDURE": TokenType.PROCEDURE,
 760        "QUALIFY": TokenType.QUALIFY,
 761        "RANGE": TokenType.RANGE,
 762        "RECURSIVE": TokenType.RECURSIVE,
 763        "REGEXP": TokenType.RLIKE,
 764        "RENAME": TokenType.RENAME,
 765        "REPLACE": TokenType.REPLACE,
 766        "RETURNING": TokenType.RETURNING,
 767        "REFERENCES": TokenType.REFERENCES,
 768        "RIGHT": TokenType.RIGHT,
 769        "RLIKE": TokenType.RLIKE,
 770        "ROLLBACK": TokenType.ROLLBACK,
 771        "ROLLUP": TokenType.ROLLUP,
 772        "ROW": TokenType.ROW,
 773        "ROWS": TokenType.ROWS,
 774        "SCHEMA": TokenType.SCHEMA,
 775        "SELECT": TokenType.SELECT,
 776        "SEMI": TokenType.SEMI,
 777        "SET": TokenType.SET,
 778        "SETTINGS": TokenType.SETTINGS,
 779        "SHOW": TokenType.SHOW,
 780        "SIMILAR TO": TokenType.SIMILAR_TO,
 781        "SOME": TokenType.SOME,
 782        "SORT BY": TokenType.SORT_BY,
 783        "START WITH": TokenType.START_WITH,
 784        "STRAIGHT_JOIN": TokenType.STRAIGHT_JOIN,
 785        "TABLE": TokenType.TABLE,
 786        "TABLESAMPLE": TokenType.TABLE_SAMPLE,
 787        "TEMP": TokenType.TEMPORARY,
 788        "TEMPORARY": TokenType.TEMPORARY,
 789        "THEN": TokenType.THEN,
 790        "TRUE": TokenType.TRUE,
 791        "TRUNCATE": TokenType.TRUNCATE,
 792        "UNION": TokenType.UNION,
 793        "UNKNOWN": TokenType.UNKNOWN,
 794        "UNNEST": TokenType.UNNEST,
 795        "UNPIVOT": TokenType.UNPIVOT,
 796        "UPDATE": TokenType.UPDATE,
 797        "USE": TokenType.USE,
 798        "USING": TokenType.USING,
 799        "UUID": TokenType.UUID,
 800        "VALUES": TokenType.VALUES,
 801        "VIEW": TokenType.VIEW,
 802        "VOLATILE": TokenType.VOLATILE,
 803        "WHEN": TokenType.WHEN,
 804        "WHERE": TokenType.WHERE,
 805        "WINDOW": TokenType.WINDOW,
 806        "WITH": TokenType.WITH,
 807        "APPLY": TokenType.APPLY,
 808        "ARRAY": TokenType.ARRAY,
 809        "BIT": TokenType.BIT,
 810        "BOOL": TokenType.BOOLEAN,
 811        "BOOLEAN": TokenType.BOOLEAN,
 812        "BYTE": TokenType.TINYINT,
 813        "MEDIUMINT": TokenType.MEDIUMINT,
 814        "INT1": TokenType.TINYINT,
 815        "TINYINT": TokenType.TINYINT,
 816        "INT16": TokenType.SMALLINT,
 817        "SHORT": TokenType.SMALLINT,
 818        "SMALLINT": TokenType.SMALLINT,
 819        "INT128": TokenType.INT128,
 820        "HUGEINT": TokenType.INT128,
 821        "INT2": TokenType.SMALLINT,
 822        "INTEGER": TokenType.INT,
 823        "INT": TokenType.INT,
 824        "INT4": TokenType.INT,
 825        "INT32": TokenType.INT,
 826        "INT64": TokenType.BIGINT,
 827        "LONG": TokenType.BIGINT,
 828        "BIGINT": TokenType.BIGINT,
 829        "INT8": TokenType.TINYINT,
 830        "UINT": TokenType.UINT,
 831        "DEC": TokenType.DECIMAL,
 832        "DECIMAL": TokenType.DECIMAL,
 833        "DECIMAL32": TokenType.DECIMAL32,
 834        "DECIMAL64": TokenType.DECIMAL64,
 835        "DECIMAL128": TokenType.DECIMAL128,
 836        "BIGDECIMAL": TokenType.BIGDECIMAL,
 837        "BIGNUMERIC": TokenType.BIGDECIMAL,
 838        "LIST": TokenType.LIST,
 839        "MAP": TokenType.MAP,
 840        "NULLABLE": TokenType.NULLABLE,
 841        "NUMBER": TokenType.DECIMAL,
 842        "NUMERIC": TokenType.DECIMAL,
 843        "FIXED": TokenType.DECIMAL,
 844        "REAL": TokenType.FLOAT,
 845        "FLOAT": TokenType.FLOAT,
 846        "FLOAT4": TokenType.FLOAT,
 847        "FLOAT8": TokenType.DOUBLE,
 848        "DOUBLE": TokenType.DOUBLE,
 849        "DOUBLE PRECISION": TokenType.DOUBLE,
 850        "JSON": TokenType.JSON,
 851        "JSONB": TokenType.JSONB,
 852        "CHAR": TokenType.CHAR,
 853        "CHARACTER": TokenType.CHAR,
 854        "NCHAR": TokenType.NCHAR,
 855        "VARCHAR": TokenType.VARCHAR,
 856        "VARCHAR2": TokenType.VARCHAR,
 857        "NVARCHAR": TokenType.NVARCHAR,
 858        "NVARCHAR2": TokenType.NVARCHAR,
 859        "BPCHAR": TokenType.BPCHAR,
 860        "STR": TokenType.TEXT,
 861        "STRING": TokenType.TEXT,
 862        "TEXT": TokenType.TEXT,
 863        "LONGTEXT": TokenType.LONGTEXT,
 864        "MEDIUMTEXT": TokenType.MEDIUMTEXT,
 865        "TINYTEXT": TokenType.TINYTEXT,
 866        "CLOB": TokenType.TEXT,
 867        "LONGVARCHAR": TokenType.TEXT,
 868        "BINARY": TokenType.BINARY,
 869        "BLOB": TokenType.VARBINARY,
 870        "LONGBLOB": TokenType.LONGBLOB,
 871        "MEDIUMBLOB": TokenType.MEDIUMBLOB,
 872        "TINYBLOB": TokenType.TINYBLOB,
 873        "BYTEA": TokenType.VARBINARY,
 874        "VARBINARY": TokenType.VARBINARY,
 875        "TIME": TokenType.TIME,
 876        "TIMETZ": TokenType.TIMETZ,
 877        "TIMESTAMP": TokenType.TIMESTAMP,
 878        "TIMESTAMPTZ": TokenType.TIMESTAMPTZ,
 879        "TIMESTAMPLTZ": TokenType.TIMESTAMPLTZ,
 880        "TIMESTAMP_LTZ": TokenType.TIMESTAMPLTZ,
 881        "TIMESTAMPNTZ": TokenType.TIMESTAMPNTZ,
 882        "TIMESTAMP_NTZ": TokenType.TIMESTAMPNTZ,
 883        "DATE": TokenType.DATE,
 884        "DATETIME": TokenType.DATETIME,
 885        "INT4RANGE": TokenType.INT4RANGE,
 886        "INT4MULTIRANGE": TokenType.INT4MULTIRANGE,
 887        "INT8RANGE": TokenType.INT8RANGE,
 888        "INT8MULTIRANGE": TokenType.INT8MULTIRANGE,
 889        "NUMRANGE": TokenType.NUMRANGE,
 890        "NUMMULTIRANGE": TokenType.NUMMULTIRANGE,
 891        "TSRANGE": TokenType.TSRANGE,
 892        "TSMULTIRANGE": TokenType.TSMULTIRANGE,
 893        "TSTZRANGE": TokenType.TSTZRANGE,
 894        "TSTZMULTIRANGE": TokenType.TSTZMULTIRANGE,
 895        "DATERANGE": TokenType.DATERANGE,
 896        "DATEMULTIRANGE": TokenType.DATEMULTIRANGE,
 897        "UNIQUE": TokenType.UNIQUE,
 898        "VECTOR": TokenType.VECTOR,
 899        "STRUCT": TokenType.STRUCT,
 900        "SEQUENCE": TokenType.SEQUENCE,
 901        "VARIANT": TokenType.VARIANT,
 902        "ALTER": TokenType.ALTER,
 903        "ANALYZE": TokenType.COMMAND,
 904        "CALL": TokenType.COMMAND,
 905        "COMMENT": TokenType.COMMENT,
 906        "EXPLAIN": TokenType.COMMAND,
 907        "GRANT": TokenType.GRANT,
 908        "OPTIMIZE": TokenType.COMMAND,
 909        "PREPARE": TokenType.COMMAND,
 910        "VACUUM": TokenType.COMMAND,
 911        "USER-DEFINED": TokenType.USERDEFINED,
 912        "FOR VERSION": TokenType.VERSION_SNAPSHOT,
 913        "FOR TIMESTAMP": TokenType.TIMESTAMP_SNAPSHOT,
 914    }
 915
 916    WHITE_SPACE: t.Dict[t.Optional[str], TokenType] = {
 917        " ": TokenType.SPACE,
 918        "\t": TokenType.SPACE,
 919        "\n": TokenType.BREAK,
 920        "\r": TokenType.BREAK,
 921    }
 922
 923    COMMANDS = {
 924        TokenType.COMMAND,
 925        TokenType.EXECUTE,
 926        TokenType.FETCH,
 927        TokenType.SHOW,
 928        TokenType.RENAME,
 929    }
 930
 931    COMMAND_PREFIX_TOKENS = {TokenType.SEMICOLON, TokenType.BEGIN}
 932
 933    # Handle numeric literals like in hive (3L = BIGINT)
 934    NUMERIC_LITERALS: t.Dict[str, str] = {}
 935
 936    COMMENTS = ["--", ("/*", "*/")]
 937
 938    __slots__ = (
 939        "sql",
 940        "size",
 941        "tokens",
 942        "dialect",
 943        "_start",
 944        "_current",
 945        "_line",
 946        "_col",
 947        "_comments",
 948        "_char",
 949        "_end",
 950        "_peek",
 951        "_prev_token_line",
 952        "_rs_dialect_settings",
 953    )
 954
 955    def __init__(self, dialect: DialectType = None) -> None:
 956        from sqlglot.dialects import Dialect
 957
 958        self.dialect = Dialect.get_or_raise(dialect)
 959
 960        if USE_RS_TOKENIZER:
 961            self._rs_dialect_settings = RsTokenizerDialectSettings(
 962                unescaped_sequences=self.dialect.UNESCAPED_SEQUENCES,
 963                identifiers_can_start_with_digit=self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT,
 964            )
 965
 966        self.reset()
 967
 968    def reset(self) -> None:
 969        self.sql = ""
 970        self.size = 0
 971        self.tokens: t.List[Token] = []
 972        self._start = 0
 973        self._current = 0
 974        self._line = 1
 975        self._col = 0
 976        self._comments: t.List[str] = []
 977
 978        self._char = ""
 979        self._end = False
 980        self._peek = ""
 981        self._prev_token_line = -1
 982
 983    def tokenize(self, sql: str) -> t.List[Token]:
 984        """Returns a list of tokens corresponding to the SQL string `sql`."""
 985        if USE_RS_TOKENIZER:
 986            return self.tokenize_rs(sql)
 987
 988        self.reset()
 989        self.sql = sql
 990        self.size = len(sql)
 991
 992        try:
 993            self._scan()
 994        except Exception as e:
 995            start = max(self._current - 50, 0)
 996            end = min(self._current + 50, self.size - 1)
 997            context = self.sql[start:end]
 998            raise TokenError(f"Error tokenizing '{context}'") from e
 999
1000        return self.tokens
1001
1002    def _scan(self, until: t.Optional[t.Callable] = None) -> None:
1003        while self.size and not self._end:
1004            current = self._current
1005
1006            # Skip spaces here rather than iteratively calling advance() for performance reasons
1007            while current < self.size:
1008                char = self.sql[current]
1009
1010                if char.isspace() and (char == " " or char == "\t"):
1011                    current += 1
1012                else:
1013                    break
1014
1015            offset = current - self._current if current > self._current else 1
1016
1017            self._start = current
1018            self._advance(offset)
1019
1020            if not self._char.isspace():
1021                if self._char.isdigit():
1022                    self._scan_number()
1023                elif self._char in self._IDENTIFIERS:
1024                    self._scan_identifier(self._IDENTIFIERS[self._char])
1025                else:
1026                    self._scan_keywords()
1027
1028            if until and until():
1029                break
1030
1031        if self.tokens and self._comments:
1032            self.tokens[-1].comments.extend(self._comments)
1033
1034    def _chars(self, size: int) -> str:
1035        if size == 1:
1036            return self._char
1037
1038        start = self._current - 1
1039        end = start + size
1040
1041        return self.sql[start:end] if end <= self.size else ""
1042
1043    def _advance(self, i: int = 1, alnum: bool = False) -> None:
1044        if self.WHITE_SPACE.get(self._char) is TokenType.BREAK:
1045            # Ensures we don't count an extra line if we get a \r\n line break sequence
1046            if not (self._char == "\r" and self._peek == "\n"):
1047                self._col = 1
1048                self._line += 1
1049        else:
1050            self._col += i
1051
1052        self._current += i
1053        self._end = self._current >= self.size
1054        self._char = self.sql[self._current - 1]
1055        self._peek = "" if self._end else self.sql[self._current]
1056
1057        if alnum and self._char.isalnum():
1058            # Here we use local variables instead of attributes for better performance
1059            _col = self._col
1060            _current = self._current
1061            _end = self._end
1062            _peek = self._peek
1063
1064            while _peek.isalnum():
1065                _col += 1
1066                _current += 1
1067                _end = _current >= self.size
1068                _peek = "" if _end else self.sql[_current]
1069
1070            self._col = _col
1071            self._current = _current
1072            self._end = _end
1073            self._peek = _peek
1074            self._char = self.sql[_current - 1]
1075
1076    @property
1077    def _text(self) -> str:
1078        return self.sql[self._start : self._current]
1079
1080    def _add(self, token_type: TokenType, text: t.Optional[str] = None) -> None:
1081        self._prev_token_line = self._line
1082
1083        if self._comments and token_type == TokenType.SEMICOLON and self.tokens:
1084            self.tokens[-1].comments.extend(self._comments)
1085            self._comments = []
1086
1087        self.tokens.append(
1088            Token(
1089                token_type,
1090                text=self._text if text is None else text,
1091                line=self._line,
1092                col=self._col,
1093                start=self._start,
1094                end=self._current - 1,
1095                comments=self._comments,
1096            )
1097        )
1098        self._comments = []
1099
1100        # If we have either a semicolon or a begin token before the command's token, we'll parse
1101        # whatever follows the command's token as a string
1102        if (
1103            token_type in self.COMMANDS
1104            and self._peek != ";"
1105            and (len(self.tokens) == 1 or self.tokens[-2].token_type in self.COMMAND_PREFIX_TOKENS)
1106        ):
1107            start = self._current
1108            tokens = len(self.tokens)
1109            self._scan(lambda: self._peek == ";")
1110            self.tokens = self.tokens[:tokens]
1111            text = self.sql[start : self._current].strip()
1112            if text:
1113                self._add(TokenType.STRING, text)
1114
1115    def _scan_keywords(self) -> None:
1116        size = 0
1117        word = None
1118        chars = self._text
1119        char = chars
1120        prev_space = False
1121        skip = False
1122        trie = self._KEYWORD_TRIE
1123        single_token = char in self.SINGLE_TOKENS
1124
1125        while chars:
1126            if skip:
1127                result = TrieResult.PREFIX
1128            else:
1129                result, trie = in_trie(trie, char.upper())
1130
1131            if result == TrieResult.FAILED:
1132                break
1133            if result == TrieResult.EXISTS:
1134                word = chars
1135
1136            end = self._current + size
1137            size += 1
1138
1139            if end < self.size:
1140                char = self.sql[end]
1141                single_token = single_token or char in self.SINGLE_TOKENS
1142                is_space = char.isspace()
1143
1144                if not is_space or not prev_space:
1145                    if is_space:
1146                        char = " "
1147                    chars += char
1148                    prev_space = is_space
1149                    skip = False
1150                else:
1151                    skip = True
1152            else:
1153                char = ""
1154                break
1155
1156        if word:
1157            if self._scan_string(word):
1158                return
1159            if self._scan_comment(word):
1160                return
1161            if prev_space or single_token or not char:
1162                self._advance(size - 1)
1163                word = word.upper()
1164                self._add(self.KEYWORDS[word], text=word)
1165                return
1166
1167        if self._char in self.SINGLE_TOKENS:
1168            self._add(self.SINGLE_TOKENS[self._char], text=self._char)
1169            return
1170
1171        self._scan_var()
1172
1173    def _scan_comment(self, comment_start: str) -> bool:
1174        if comment_start not in self._COMMENTS:
1175            return False
1176
1177        comment_start_line = self._line
1178        comment_start_size = len(comment_start)
1179        comment_end = self._COMMENTS[comment_start]
1180
1181        if comment_end:
1182            # Skip the comment's start delimiter
1183            self._advance(comment_start_size)
1184
1185            comment_count = 1
1186            comment_end_size = len(comment_end)
1187
1188            while not self._end:
1189                if self._chars(comment_end_size) == comment_end:
1190                    comment_count -= 1
1191                    if not comment_count:
1192                        break
1193
1194                self._advance(alnum=True)
1195
1196                # Nested comments are allowed by some dialects, e.g. databricks, duckdb, postgres
1197                if (
1198                    self.NESTED_COMMENTS
1199                    and not self._end
1200                    and self._chars(comment_end_size) == comment_start
1201                ):
1202                    self._advance(comment_start_size)
1203                    comment_count += 1
1204
1205            self._comments.append(self._text[comment_start_size : -comment_end_size + 1])
1206            self._advance(comment_end_size - 1)
1207        else:
1208            while not self._end and self.WHITE_SPACE.get(self._peek) is not TokenType.BREAK:
1209                self._advance(alnum=True)
1210            self._comments.append(self._text[comment_start_size:])
1211
1212        # Leading comment is attached to the succeeding token, whilst trailing comment to the preceding.
1213        # Multiple consecutive comments are preserved by appending them to the current comments list.
1214        if comment_start_line == self._prev_token_line:
1215            self.tokens[-1].comments.extend(self._comments)
1216            self._comments = []
1217            self._prev_token_line = self._line
1218
1219        return True
1220
1221    def _scan_number(self) -> None:
1222        if self._char == "0":
1223            peek = self._peek.upper()
1224            if peek == "B":
1225                return self._scan_bits() if self.BIT_STRINGS else self._add(TokenType.NUMBER)
1226            elif peek == "X":
1227                return self._scan_hex() if self.HEX_STRINGS else self._add(TokenType.NUMBER)
1228
1229        decimal = False
1230        scientific = 0
1231
1232        while True:
1233            if self._peek.isdigit():
1234                self._advance()
1235            elif self._peek == "." and not decimal:
1236                if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER:
1237                    return self._add(TokenType.NUMBER)
1238                decimal = True
1239                self._advance()
1240            elif self._peek in ("-", "+") and scientific == 1:
1241                scientific += 1
1242                self._advance()
1243            elif self._peek.upper() == "E" and not scientific:
1244                scientific += 1
1245                self._advance()
1246            elif self._peek.isidentifier():
1247                number_text = self._text
1248                literal = ""
1249
1250                while self._peek.strip() and self._peek not in self.SINGLE_TOKENS:
1251                    literal += self._peek
1252                    self._advance()
1253
1254                token_type = self.KEYWORDS.get(self.NUMERIC_LITERALS.get(literal.upper(), ""))
1255
1256                if token_type:
1257                    self._add(TokenType.NUMBER, number_text)
1258                    self._add(TokenType.DCOLON, "::")
1259                    return self._add(token_type, literal)
1260                elif self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT:
1261                    return self._add(TokenType.VAR)
1262
1263                self._advance(-len(literal))
1264                return self._add(TokenType.NUMBER, number_text)
1265            else:
1266                return self._add(TokenType.NUMBER)
1267
1268    def _scan_bits(self) -> None:
1269        self._advance()
1270        value = self._extract_value()
1271        try:
1272            # If `value` can't be converted to a binary, fallback to tokenizing it as an identifier
1273            int(value, 2)
1274            self._add(TokenType.BIT_STRING, value[2:])  # Drop the 0b
1275        except ValueError:
1276            self._add(TokenType.IDENTIFIER)
1277
1278    def _scan_hex(self) -> None:
1279        self._advance()
1280        value = self._extract_value()
1281        try:
1282            # If `value` can't be converted to a hex, fallback to tokenizing it as an identifier
1283            int(value, 16)
1284            self._add(TokenType.HEX_STRING, value[2:])  # Drop the 0x
1285        except ValueError:
1286            self._add(TokenType.IDENTIFIER)
1287
1288    def _extract_value(self) -> str:
1289        while True:
1290            char = self._peek.strip()
1291            if char and char not in self.SINGLE_TOKENS:
1292                self._advance(alnum=True)
1293            else:
1294                break
1295
1296        return self._text
1297
1298    def _scan_string(self, start: str) -> bool:
1299        base = None
1300        token_type = TokenType.STRING
1301
1302        if start in self._QUOTES:
1303            end = self._QUOTES[start]
1304        elif start in self._FORMAT_STRINGS:
1305            end, token_type = self._FORMAT_STRINGS[start]
1306
1307            if token_type == TokenType.HEX_STRING:
1308                base = 16
1309            elif token_type == TokenType.BIT_STRING:
1310                base = 2
1311            elif token_type == TokenType.HEREDOC_STRING:
1312                self._advance()
1313
1314                if self._char == end:
1315                    tag = ""
1316                else:
1317                    tag = self._extract_string(
1318                        end,
1319                        raw_string=True,
1320                        raise_unmatched=not self.HEREDOC_TAG_IS_IDENTIFIER,
1321                    )
1322
1323                if tag and self.HEREDOC_TAG_IS_IDENTIFIER and (self._end or not tag.isidentifier()):
1324                    if not self._end:
1325                        self._advance(-1)
1326
1327                    self._advance(-len(tag))
1328                    self._add(self.HEREDOC_STRING_ALTERNATIVE)
1329                    return True
1330
1331                end = f"{start}{tag}{end}"
1332        else:
1333            return False
1334
1335        self._advance(len(start))
1336        text = self._extract_string(end, raw_string=token_type == TokenType.RAW_STRING)
1337
1338        if base:
1339            try:
1340                int(text, base)
1341            except Exception:
1342                raise TokenError(
1343                    f"Numeric string contains invalid characters from {self._line}:{self._start}"
1344                )
1345
1346        self._add(token_type, text)
1347        return True
1348
1349    def _scan_identifier(self, identifier_end: str) -> None:
1350        self._advance()
1351        text = self._extract_string(identifier_end, escapes=self._IDENTIFIER_ESCAPES)
1352        self._add(TokenType.IDENTIFIER, text)
1353
1354    def _scan_var(self) -> None:
1355        while True:
1356            char = self._peek.strip()
1357            if char and (char in self.VAR_SINGLE_TOKENS or char not in self.SINGLE_TOKENS):
1358                self._advance(alnum=True)
1359            else:
1360                break
1361
1362        self._add(
1363            TokenType.VAR
1364            if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER
1365            else self.KEYWORDS.get(self._text.upper(), TokenType.VAR)
1366        )
1367
1368    def _extract_string(
1369        self,
1370        delimiter: str,
1371        escapes: t.Optional[t.Set[str]] = None,
1372        raw_string: bool = False,
1373        raise_unmatched: bool = True,
1374    ) -> str:
1375        text = ""
1376        delim_size = len(delimiter)
1377        escapes = self._STRING_ESCAPES if escapes is None else escapes
1378
1379        while True:
1380            if (
1381                not raw_string
1382                and self.dialect.UNESCAPED_SEQUENCES
1383                and self._peek
1384                and self._char in self.STRING_ESCAPES
1385            ):
1386                unescaped_sequence = self.dialect.UNESCAPED_SEQUENCES.get(self._char + self._peek)
1387                if unescaped_sequence:
1388                    self._advance(2)
1389                    text += unescaped_sequence
1390                    continue
1391            if (
1392                (self.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS or not raw_string)
1393                and self._char in escapes
1394                and (self._peek == delimiter or self._peek in escapes)
1395                and (self._char not in self._QUOTES or self._char == self._peek)
1396            ):
1397                if self._peek == delimiter:
1398                    text += self._peek
1399                else:
1400                    text += self._char + self._peek
1401
1402                if self._current + 1 < self.size:
1403                    self._advance(2)
1404                else:
1405                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._current}")
1406            else:
1407                if self._chars(delim_size) == delimiter:
1408                    if delim_size > 1:
1409                        self._advance(delim_size - 1)
1410                    break
1411
1412                if self._end:
1413                    if not raise_unmatched:
1414                        return text + self._char
1415
1416                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._start}")
1417
1418                current = self._current - 1
1419                self._advance(alnum=True)
1420                text += self.sql[current : self._current - 1]
1421
1422        return text
1423
1424    def tokenize_rs(self, sql: str) -> t.List[Token]:
1425        if not self._RS_TOKENIZER:
1426            raise SqlglotError("Rust tokenizer is not available")
1427
1428        try:
1429            tokens = self._RS_TOKENIZER.tokenize(sql, self._rs_dialect_settings)
1430            for token in tokens:
1431                token.token_type = _ALL_TOKEN_TYPES[token.token_type_index]
1432            return tokens
1433        except Exception as e:
1434            raise TokenError(str(e))