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    CARET_AT = auto()
  64    TILDA = auto()
  65    ARROW = auto()
  66    DARROW = auto()
  67    FARROW = auto()
  68    HASH = auto()
  69    HASH_ARROW = auto()
  70    DHASH_ARROW = auto()
  71    LR_ARROW = auto()
  72    DAT = auto()
  73    LT_AT = auto()
  74    AT_GT = auto()
  75    DOLLAR = auto()
  76    PARAMETER = auto()
  77    SESSION_PARAMETER = auto()
  78    DAMP = auto()
  79    XOR = auto()
  80    DSTAR = auto()
  81
  82    BLOCK_START = auto()
  83    BLOCK_END = auto()
  84
  85    SPACE = auto()
  86    BREAK = auto()
  87
  88    STRING = auto()
  89    NUMBER = auto()
  90    IDENTIFIER = auto()
  91    DATABASE = auto()
  92    COLUMN = auto()
  93    COLUMN_DEF = auto()
  94    SCHEMA = auto()
  95    TABLE = auto()
  96    WAREHOUSE = auto()
  97    STREAMLIT = auto()
  98    VAR = auto()
  99    BIT_STRING = auto()
 100    HEX_STRING = auto()
 101    BYTE_STRING = auto()
 102    NATIONAL_STRING = auto()
 103    RAW_STRING = auto()
 104    HEREDOC_STRING = auto()
 105    UNICODE_STRING = auto()
 106
 107    # types
 108    BIT = auto()
 109    BOOLEAN = auto()
 110    TINYINT = auto()
 111    UTINYINT = auto()
 112    SMALLINT = auto()
 113    USMALLINT = auto()
 114    MEDIUMINT = auto()
 115    UMEDIUMINT = auto()
 116    INT = auto()
 117    UINT = auto()
 118    BIGINT = auto()
 119    UBIGINT = auto()
 120    INT128 = auto()
 121    UINT128 = auto()
 122    INT256 = auto()
 123    UINT256 = auto()
 124    FLOAT = auto()
 125    DOUBLE = auto()
 126    DECIMAL = auto()
 127    DECIMAL32 = auto()
 128    DECIMAL64 = auto()
 129    DECIMAL128 = auto()
 130    DECIMAL256 = auto()
 131    UDECIMAL = auto()
 132    BIGDECIMAL = auto()
 133    CHAR = auto()
 134    NCHAR = auto()
 135    VARCHAR = auto()
 136    NVARCHAR = auto()
 137    BPCHAR = auto()
 138    TEXT = auto()
 139    MEDIUMTEXT = auto()
 140    LONGTEXT = auto()
 141    MEDIUMBLOB = auto()
 142    LONGBLOB = auto()
 143    TINYBLOB = auto()
 144    TINYTEXT = auto()
 145    NAME = auto()
 146    BINARY = auto()
 147    VARBINARY = auto()
 148    JSON = auto()
 149    JSONB = auto()
 150    TIME = auto()
 151    TIMETZ = auto()
 152    TIMESTAMP = auto()
 153    TIMESTAMPTZ = auto()
 154    TIMESTAMPLTZ = auto()
 155    TIMESTAMPNTZ = auto()
 156    TIMESTAMP_S = auto()
 157    TIMESTAMP_MS = auto()
 158    TIMESTAMP_NS = auto()
 159    DATETIME = auto()
 160    DATETIME64 = auto()
 161    DATE = auto()
 162    DATE32 = auto()
 163    INT4RANGE = auto()
 164    INT4MULTIRANGE = auto()
 165    INT8RANGE = auto()
 166    INT8MULTIRANGE = auto()
 167    NUMRANGE = auto()
 168    NUMMULTIRANGE = auto()
 169    TSRANGE = auto()
 170    TSMULTIRANGE = auto()
 171    TSTZRANGE = auto()
 172    TSTZMULTIRANGE = auto()
 173    DATERANGE = auto()
 174    DATEMULTIRANGE = auto()
 175    UUID = auto()
 176    GEOGRAPHY = auto()
 177    NULLABLE = auto()
 178    GEOMETRY = auto()
 179    POINT = auto()
 180    RING = auto()
 181    LINESTRING = auto()
 182    MULTILINESTRING = auto()
 183    POLYGON = auto()
 184    MULTIPOLYGON = auto()
 185    HLLSKETCH = auto()
 186    HSTORE = auto()
 187    SUPER = auto()
 188    SERIAL = auto()
 189    SMALLSERIAL = auto()
 190    BIGSERIAL = auto()
 191    XML = auto()
 192    YEAR = auto()
 193    UNIQUEIDENTIFIER = auto()
 194    USERDEFINED = auto()
 195    MONEY = auto()
 196    SMALLMONEY = auto()
 197    ROWVERSION = auto()
 198    IMAGE = auto()
 199    VARIANT = auto()
 200    OBJECT = auto()
 201    INET = auto()
 202    IPADDRESS = auto()
 203    IPPREFIX = auto()
 204    IPV4 = auto()
 205    IPV6 = auto()
 206    ENUM = auto()
 207    ENUM8 = auto()
 208    ENUM16 = auto()
 209    FIXEDSTRING = auto()
 210    LOWCARDINALITY = auto()
 211    NESTED = auto()
 212    AGGREGATEFUNCTION = auto()
 213    SIMPLEAGGREGATEFUNCTION = auto()
 214    TDIGEST = auto()
 215    UNKNOWN = auto()
 216    VECTOR = auto()
 217
 218    # keywords
 219    ALIAS = auto()
 220    ALTER = auto()
 221    ALWAYS = auto()
 222    ALL = auto()
 223    ANTI = auto()
 224    ANY = auto()
 225    APPLY = auto()
 226    ARRAY = auto()
 227    ASC = auto()
 228    ASOF = auto()
 229    AUTO_INCREMENT = auto()
 230    BEGIN = auto()
 231    BETWEEN = auto()
 232    BULK_COLLECT_INTO = auto()
 233    CACHE = auto()
 234    CASE = auto()
 235    CHARACTER_SET = auto()
 236    CLUSTER_BY = auto()
 237    COLLATE = auto()
 238    COMMAND = auto()
 239    COMMENT = auto()
 240    COMMIT = auto()
 241    CONNECT_BY = auto()
 242    CONSTRAINT = auto()
 243    COPY = auto()
 244    CREATE = auto()
 245    CROSS = auto()
 246    CUBE = auto()
 247    CURRENT_DATE = auto()
 248    CURRENT_DATETIME = auto()
 249    CURRENT_TIME = auto()
 250    CURRENT_TIMESTAMP = auto()
 251    CURRENT_USER = auto()
 252    DECLARE = auto()
 253    DEFAULT = auto()
 254    DELETE = auto()
 255    DESC = auto()
 256    DESCRIBE = auto()
 257    DICTIONARY = auto()
 258    DISTINCT = auto()
 259    DISTRIBUTE_BY = auto()
 260    DIV = auto()
 261    DROP = auto()
 262    ELSE = auto()
 263    END = auto()
 264    ESCAPE = auto()
 265    EXCEPT = auto()
 266    EXECUTE = auto()
 267    EXISTS = auto()
 268    FALSE = auto()
 269    FETCH = auto()
 270    FILTER = auto()
 271    FINAL = auto()
 272    FIRST = auto()
 273    FOR = auto()
 274    FORCE = auto()
 275    FOREIGN_KEY = auto()
 276    FORMAT = auto()
 277    FROM = auto()
 278    FULL = auto()
 279    FUNCTION = auto()
 280    GLOB = auto()
 281    GLOBAL = auto()
 282    GRANT = auto()
 283    GROUP_BY = auto()
 284    GROUPING_SETS = auto()
 285    HAVING = auto()
 286    HINT = auto()
 287    IGNORE = auto()
 288    ILIKE = auto()
 289    ILIKE_ANY = auto()
 290    IN = auto()
 291    INDEX = auto()
 292    INNER = auto()
 293    INSERT = auto()
 294    INTERSECT = auto()
 295    INTERVAL = auto()
 296    INTO = auto()
 297    INTRODUCER = auto()
 298    IRLIKE = auto()
 299    IS = auto()
 300    ISNULL = auto()
 301    JOIN = auto()
 302    JOIN_MARKER = auto()
 303    KEEP = auto()
 304    KEY = auto()
 305    KILL = auto()
 306    LANGUAGE = auto()
 307    LATERAL = auto()
 308    LEFT = auto()
 309    LIKE = auto()
 310    LIKE_ANY = auto()
 311    LIMIT = auto()
 312    LIST = auto()
 313    LOAD = auto()
 314    LOCK = auto()
 315    MAP = auto()
 316    MATCH_CONDITION = auto()
 317    MATCH_RECOGNIZE = auto()
 318    MEMBER_OF = auto()
 319    MERGE = auto()
 320    MOD = auto()
 321    MODEL = auto()
 322    NATURAL = auto()
 323    NEXT = auto()
 324    NOTNULL = auto()
 325    NULL = auto()
 326    OBJECT_IDENTIFIER = auto()
 327    OFFSET = auto()
 328    ON = auto()
 329    ONLY = auto()
 330    OPERATOR = auto()
 331    ORDER_BY = auto()
 332    ORDER_SIBLINGS_BY = auto()
 333    ORDERED = auto()
 334    ORDINALITY = auto()
 335    OUTER = auto()
 336    OVER = auto()
 337    OVERLAPS = auto()
 338    OVERWRITE = auto()
 339    PARTITION = auto()
 340    PARTITION_BY = auto()
 341    PERCENT = auto()
 342    PIVOT = auto()
 343    PLACEHOLDER = auto()
 344    POSITIONAL = auto()
 345    PRAGMA = auto()
 346    PREWHERE = auto()
 347    PRIMARY_KEY = auto()
 348    PROCEDURE = auto()
 349    PROPERTIES = auto()
 350    PSEUDO_TYPE = auto()
 351    QUALIFY = auto()
 352    QUOTE = auto()
 353    RANGE = auto()
 354    RECURSIVE = auto()
 355    REFRESH = auto()
 356    RENAME = auto()
 357    REPLACE = auto()
 358    RETURNING = auto()
 359    REFERENCES = auto()
 360    RIGHT = auto()
 361    RLIKE = auto()
 362    ROLLBACK = auto()
 363    ROLLUP = auto()
 364    ROW = auto()
 365    ROWS = auto()
 366    SELECT = auto()
 367    SEMI = auto()
 368    SEPARATOR = auto()
 369    SEQUENCE = auto()
 370    SERDE_PROPERTIES = auto()
 371    SET = auto()
 372    SETTINGS = auto()
 373    SHOW = auto()
 374    SIMILAR_TO = auto()
 375    SOME = auto()
 376    SORT_BY = auto()
 377    START_WITH = auto()
 378    STORAGE_INTEGRATION = auto()
 379    STRAIGHT_JOIN = auto()
 380    STRUCT = auto()
 381    SUMMARIZE = auto()
 382    TABLE_SAMPLE = auto()
 383    TAG = auto()
 384    TEMPORARY = auto()
 385    TOP = auto()
 386    THEN = auto()
 387    TRUE = auto()
 388    TRUNCATE = auto()
 389    UNCACHE = auto()
 390    UNION = auto()
 391    UNNEST = auto()
 392    UNPIVOT = auto()
 393    UPDATE = auto()
 394    USE = auto()
 395    USING = auto()
 396    VALUES = auto()
 397    VIEW = auto()
 398    VOLATILE = auto()
 399    WHEN = auto()
 400    WHERE = auto()
 401    WINDOW = auto()
 402    WITH = auto()
 403    UNIQUE = auto()
 404    VERSION_SNAPSHOT = auto()
 405    TIMESTAMP_SNAPSHOT = auto()
 406    OPTION = auto()
 407
 408
 409_ALL_TOKEN_TYPES = list(TokenType)
 410_TOKEN_TYPE_TO_INDEX = {token_type: i for i, token_type in enumerate(_ALL_TOKEN_TYPES)}
 411
 412
 413class Token:
 414    __slots__ = ("token_type", "text", "line", "col", "start", "end", "comments")
 415
 416    @classmethod
 417    def number(cls, number: int) -> Token:
 418        """Returns a NUMBER token with `number` as its text."""
 419        return cls(TokenType.NUMBER, str(number))
 420
 421    @classmethod
 422    def string(cls, string: str) -> Token:
 423        """Returns a STRING token with `string` as its text."""
 424        return cls(TokenType.STRING, string)
 425
 426    @classmethod
 427    def identifier(cls, identifier: str) -> Token:
 428        """Returns an IDENTIFIER token with `identifier` as its text."""
 429        return cls(TokenType.IDENTIFIER, identifier)
 430
 431    @classmethod
 432    def var(cls, var: str) -> Token:
 433        """Returns an VAR token with `var` as its text."""
 434        return cls(TokenType.VAR, var)
 435
 436    def __init__(
 437        self,
 438        token_type: TokenType,
 439        text: str,
 440        line: int = 1,
 441        col: int = 1,
 442        start: int = 0,
 443        end: int = 0,
 444        comments: t.Optional[t.List[str]] = None,
 445    ) -> None:
 446        """Token initializer.
 447
 448        Args:
 449            token_type: The TokenType Enum.
 450            text: The text of the token.
 451            line: The line that the token ends on.
 452            col: The column that the token ends on.
 453            start: The start index of the token.
 454            end: The ending index of the token.
 455            comments: The comments to attach to the token.
 456        """
 457        self.token_type = token_type
 458        self.text = text
 459        self.line = line
 460        self.col = col
 461        self.start = start
 462        self.end = end
 463        self.comments = [] if comments is None else comments
 464
 465    def __repr__(self) -> str:
 466        attributes = ", ".join(f"{k}: {getattr(self, k)}" for k in self.__slots__)
 467        return f"<Token {attributes}>"
 468
 469
 470class _Tokenizer(type):
 471    def __new__(cls, clsname, bases, attrs):
 472        klass = super().__new__(cls, clsname, bases, attrs)
 473
 474        def _convert_quotes(arr: t.List[str | t.Tuple[str, str]]) -> t.Dict[str, str]:
 475            return dict(
 476                (item, item) if isinstance(item, str) else (item[0], item[1]) for item in arr
 477            )
 478
 479        def _quotes_to_format(
 480            token_type: TokenType, arr: t.List[str | t.Tuple[str, str]]
 481        ) -> t.Dict[str, t.Tuple[str, TokenType]]:
 482            return {k: (v, token_type) for k, v in _convert_quotes(arr).items()}
 483
 484        klass._QUOTES = _convert_quotes(klass.QUOTES)
 485        klass._IDENTIFIERS = _convert_quotes(klass.IDENTIFIERS)
 486
 487        klass._FORMAT_STRINGS = {
 488            **{
 489                p + s: (e, TokenType.NATIONAL_STRING)
 490                for s, e in klass._QUOTES.items()
 491                for p in ("n", "N")
 492            },
 493            **_quotes_to_format(TokenType.BIT_STRING, klass.BIT_STRINGS),
 494            **_quotes_to_format(TokenType.BYTE_STRING, klass.BYTE_STRINGS),
 495            **_quotes_to_format(TokenType.HEX_STRING, klass.HEX_STRINGS),
 496            **_quotes_to_format(TokenType.RAW_STRING, klass.RAW_STRINGS),
 497            **_quotes_to_format(TokenType.HEREDOC_STRING, klass.HEREDOC_STRINGS),
 498            **_quotes_to_format(TokenType.UNICODE_STRING, klass.UNICODE_STRINGS),
 499        }
 500
 501        klass._STRING_ESCAPES = set(klass.STRING_ESCAPES)
 502        klass._IDENTIFIER_ESCAPES = set(klass.IDENTIFIER_ESCAPES)
 503        klass._COMMENTS = {
 504            **dict(
 505                (comment, None) if isinstance(comment, str) else (comment[0], comment[1])
 506                for comment in klass.COMMENTS
 507            ),
 508            "{#": "#}",  # Ensure Jinja comments are tokenized correctly in all dialects
 509        }
 510
 511        klass._KEYWORD_TRIE = new_trie(
 512            key.upper()
 513            for key in (
 514                *klass.KEYWORDS,
 515                *klass._COMMENTS,
 516                *klass._QUOTES,
 517                *klass._FORMAT_STRINGS,
 518            )
 519            if " " in key or any(single in key for single in klass.SINGLE_TOKENS)
 520        )
 521
 522        if USE_RS_TOKENIZER:
 523            settings = RsTokenizerSettings(
 524                white_space={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.WHITE_SPACE.items()},
 525                single_tokens={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.SINGLE_TOKENS.items()},
 526                keywords={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.KEYWORDS.items()},
 527                numeric_literals=klass.NUMERIC_LITERALS,
 528                identifiers=klass._IDENTIFIERS,
 529                identifier_escapes=klass._IDENTIFIER_ESCAPES,
 530                string_escapes=klass._STRING_ESCAPES,
 531                quotes=klass._QUOTES,
 532                format_strings={
 533                    k: (v1, _TOKEN_TYPE_TO_INDEX[v2])
 534                    for k, (v1, v2) in klass._FORMAT_STRINGS.items()
 535                },
 536                has_bit_strings=bool(klass.BIT_STRINGS),
 537                has_hex_strings=bool(klass.HEX_STRINGS),
 538                comments=klass._COMMENTS,
 539                var_single_tokens=klass.VAR_SINGLE_TOKENS,
 540                commands={_TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMANDS},
 541                command_prefix_tokens={
 542                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMAND_PREFIX_TOKENS
 543                },
 544                heredoc_tag_is_identifier=klass.HEREDOC_TAG_IS_IDENTIFIER,
 545                string_escapes_allowed_in_raw_strings=klass.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS,
 546                nested_comments=klass.NESTED_COMMENTS,
 547            )
 548            token_types = RsTokenTypeSettings(
 549                bit_string=_TOKEN_TYPE_TO_INDEX[TokenType.BIT_STRING],
 550                break_=_TOKEN_TYPE_TO_INDEX[TokenType.BREAK],
 551                dcolon=_TOKEN_TYPE_TO_INDEX[TokenType.DCOLON],
 552                heredoc_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEREDOC_STRING],
 553                raw_string=_TOKEN_TYPE_TO_INDEX[TokenType.RAW_STRING],
 554                hex_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEX_STRING],
 555                identifier=_TOKEN_TYPE_TO_INDEX[TokenType.IDENTIFIER],
 556                number=_TOKEN_TYPE_TO_INDEX[TokenType.NUMBER],
 557                parameter=_TOKEN_TYPE_TO_INDEX[TokenType.PARAMETER],
 558                semicolon=_TOKEN_TYPE_TO_INDEX[TokenType.SEMICOLON],
 559                string=_TOKEN_TYPE_TO_INDEX[TokenType.STRING],
 560                var=_TOKEN_TYPE_TO_INDEX[TokenType.VAR],
 561                heredoc_string_alternative=_TOKEN_TYPE_TO_INDEX[klass.HEREDOC_STRING_ALTERNATIVE],
 562            )
 563            klass._RS_TOKENIZER = RsTokenizer(settings, token_types)
 564        else:
 565            klass._RS_TOKENIZER = None
 566
 567        return klass
 568
 569
 570class Tokenizer(metaclass=_Tokenizer):
 571    SINGLE_TOKENS = {
 572        "(": TokenType.L_PAREN,
 573        ")": TokenType.R_PAREN,
 574        "[": TokenType.L_BRACKET,
 575        "]": TokenType.R_BRACKET,
 576        "{": TokenType.L_BRACE,
 577        "}": TokenType.R_BRACE,
 578        "&": TokenType.AMP,
 579        "^": TokenType.CARET,
 580        ":": TokenType.COLON,
 581        ",": TokenType.COMMA,
 582        ".": TokenType.DOT,
 583        "-": TokenType.DASH,
 584        "=": TokenType.EQ,
 585        ">": TokenType.GT,
 586        "<": TokenType.LT,
 587        "%": TokenType.MOD,
 588        "!": TokenType.NOT,
 589        "|": TokenType.PIPE,
 590        "+": TokenType.PLUS,
 591        ";": TokenType.SEMICOLON,
 592        "/": TokenType.SLASH,
 593        "\\": TokenType.BACKSLASH,
 594        "*": TokenType.STAR,
 595        "~": TokenType.TILDA,
 596        "?": TokenType.PLACEHOLDER,
 597        "@": TokenType.PARAMETER,
 598        "#": TokenType.HASH,
 599        # Used for breaking a var like x'y' but nothing else the token type doesn't matter
 600        "'": TokenType.UNKNOWN,
 601        "`": TokenType.UNKNOWN,
 602        '"': TokenType.UNKNOWN,
 603    }
 604
 605    BIT_STRINGS: t.List[str | t.Tuple[str, str]] = []
 606    BYTE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 607    HEX_STRINGS: t.List[str | t.Tuple[str, str]] = []
 608    RAW_STRINGS: t.List[str | t.Tuple[str, str]] = []
 609    HEREDOC_STRINGS: t.List[str | t.Tuple[str, str]] = []
 610    UNICODE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 611    IDENTIFIERS: t.List[str | t.Tuple[str, str]] = ['"']
 612    QUOTES: t.List[t.Tuple[str, str] | str] = ["'"]
 613    STRING_ESCAPES = ["'"]
 614    VAR_SINGLE_TOKENS: t.Set[str] = set()
 615
 616    # The strings in this list can always be used as escapes, regardless of the surrounding
 617    # identifier delimiters. By default, the closing delimiter is assumed to also act as an
 618    # identifier escape, e.g. if we use double-quotes, then they also act as escapes: "x"""
 619    IDENTIFIER_ESCAPES: t.List[str] = []
 620
 621    # Whether the heredoc tags follow the same lexical rules as unquoted identifiers
 622    HEREDOC_TAG_IS_IDENTIFIER = False
 623
 624    # Token that we'll generate as a fallback if the heredoc prefix doesn't correspond to a heredoc
 625    HEREDOC_STRING_ALTERNATIVE = TokenType.VAR
 626
 627    # Whether string escape characters function as such when placed within raw strings
 628    STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS = True
 629
 630    NESTED_COMMENTS = True
 631
 632    # Autofilled
 633    _COMMENTS: t.Dict[str, str] = {}
 634    _FORMAT_STRINGS: t.Dict[str, t.Tuple[str, TokenType]] = {}
 635    _IDENTIFIERS: t.Dict[str, str] = {}
 636    _IDENTIFIER_ESCAPES: t.Set[str] = set()
 637    _QUOTES: t.Dict[str, str] = {}
 638    _STRING_ESCAPES: t.Set[str] = set()
 639    _KEYWORD_TRIE: t.Dict = {}
 640    _RS_TOKENIZER: t.Optional[t.Any] = None
 641
 642    KEYWORDS: t.Dict[str, TokenType] = {
 643        **{f"{{%{postfix}": TokenType.BLOCK_START for postfix in ("", "+", "-")},
 644        **{f"{prefix}%}}": TokenType.BLOCK_END for prefix in ("", "+", "-")},
 645        **{f"{{{{{postfix}": TokenType.BLOCK_START for postfix in ("+", "-")},
 646        **{f"{prefix}}}}}": TokenType.BLOCK_END for prefix in ("+", "-")},
 647        "/*+": TokenType.HINT,
 648        "==": TokenType.EQ,
 649        "::": TokenType.DCOLON,
 650        "||": TokenType.DPIPE,
 651        ">=": TokenType.GTE,
 652        "<=": TokenType.LTE,
 653        "<>": TokenType.NEQ,
 654        "!=": TokenType.NEQ,
 655        ":=": TokenType.COLON_EQ,
 656        "<=>": TokenType.NULLSAFE_EQ,
 657        "->": TokenType.ARROW,
 658        "->>": TokenType.DARROW,
 659        "=>": TokenType.FARROW,
 660        "#>": TokenType.HASH_ARROW,
 661        "#>>": TokenType.DHASH_ARROW,
 662        "<->": TokenType.LR_ARROW,
 663        "&&": TokenType.DAMP,
 664        "??": TokenType.DQMARK,
 665        "~~~": TokenType.GLOB,
 666        "~~": TokenType.LIKE,
 667        "~~*": TokenType.ILIKE,
 668        "~*": TokenType.IRLIKE,
 669        "ALL": TokenType.ALL,
 670        "ALWAYS": TokenType.ALWAYS,
 671        "AND": TokenType.AND,
 672        "ANTI": TokenType.ANTI,
 673        "ANY": TokenType.ANY,
 674        "ASC": TokenType.ASC,
 675        "AS": TokenType.ALIAS,
 676        "ASOF": TokenType.ASOF,
 677        "AUTOINCREMENT": TokenType.AUTO_INCREMENT,
 678        "AUTO_INCREMENT": TokenType.AUTO_INCREMENT,
 679        "BEGIN": TokenType.BEGIN,
 680        "BETWEEN": TokenType.BETWEEN,
 681        "CACHE": TokenType.CACHE,
 682        "UNCACHE": TokenType.UNCACHE,
 683        "CASE": TokenType.CASE,
 684        "CHARACTER SET": TokenType.CHARACTER_SET,
 685        "CLUSTER BY": TokenType.CLUSTER_BY,
 686        "COLLATE": TokenType.COLLATE,
 687        "COLUMN": TokenType.COLUMN,
 688        "COMMIT": TokenType.COMMIT,
 689        "CONNECT BY": TokenType.CONNECT_BY,
 690        "CONSTRAINT": TokenType.CONSTRAINT,
 691        "COPY": TokenType.COPY,
 692        "CREATE": TokenType.CREATE,
 693        "CROSS": TokenType.CROSS,
 694        "CUBE": TokenType.CUBE,
 695        "CURRENT_DATE": TokenType.CURRENT_DATE,
 696        "CURRENT_TIME": TokenType.CURRENT_TIME,
 697        "CURRENT_TIMESTAMP": TokenType.CURRENT_TIMESTAMP,
 698        "CURRENT_USER": TokenType.CURRENT_USER,
 699        "DATABASE": TokenType.DATABASE,
 700        "DEFAULT": TokenType.DEFAULT,
 701        "DELETE": TokenType.DELETE,
 702        "DESC": TokenType.DESC,
 703        "DESCRIBE": TokenType.DESCRIBE,
 704        "DISTINCT": TokenType.DISTINCT,
 705        "DISTRIBUTE BY": TokenType.DISTRIBUTE_BY,
 706        "DIV": TokenType.DIV,
 707        "DROP": TokenType.DROP,
 708        "ELSE": TokenType.ELSE,
 709        "END": TokenType.END,
 710        "ENUM": TokenType.ENUM,
 711        "ESCAPE": TokenType.ESCAPE,
 712        "EXCEPT": TokenType.EXCEPT,
 713        "EXECUTE": TokenType.EXECUTE,
 714        "EXISTS": TokenType.EXISTS,
 715        "FALSE": TokenType.FALSE,
 716        "FETCH": TokenType.FETCH,
 717        "FILTER": TokenType.FILTER,
 718        "FIRST": TokenType.FIRST,
 719        "FULL": TokenType.FULL,
 720        "FUNCTION": TokenType.FUNCTION,
 721        "FOR": TokenType.FOR,
 722        "FOREIGN KEY": TokenType.FOREIGN_KEY,
 723        "FORMAT": TokenType.FORMAT,
 724        "FROM": TokenType.FROM,
 725        "GEOGRAPHY": TokenType.GEOGRAPHY,
 726        "GEOMETRY": TokenType.GEOMETRY,
 727        "GLOB": TokenType.GLOB,
 728        "GROUP BY": TokenType.GROUP_BY,
 729        "GROUPING SETS": TokenType.GROUPING_SETS,
 730        "HAVING": TokenType.HAVING,
 731        "ILIKE": TokenType.ILIKE,
 732        "IN": TokenType.IN,
 733        "INDEX": TokenType.INDEX,
 734        "INET": TokenType.INET,
 735        "INNER": TokenType.INNER,
 736        "INSERT": TokenType.INSERT,
 737        "INTERVAL": TokenType.INTERVAL,
 738        "INTERSECT": TokenType.INTERSECT,
 739        "INTO": TokenType.INTO,
 740        "IS": TokenType.IS,
 741        "ISNULL": TokenType.ISNULL,
 742        "JOIN": TokenType.JOIN,
 743        "KEEP": TokenType.KEEP,
 744        "KILL": TokenType.KILL,
 745        "LATERAL": TokenType.LATERAL,
 746        "LEFT": TokenType.LEFT,
 747        "LIKE": TokenType.LIKE,
 748        "LIMIT": TokenType.LIMIT,
 749        "LOAD": TokenType.LOAD,
 750        "LOCK": TokenType.LOCK,
 751        "MERGE": TokenType.MERGE,
 752        "NATURAL": TokenType.NATURAL,
 753        "NEXT": TokenType.NEXT,
 754        "NOT": TokenType.NOT,
 755        "NOTNULL": TokenType.NOTNULL,
 756        "NULL": TokenType.NULL,
 757        "OBJECT": TokenType.OBJECT,
 758        "OFFSET": TokenType.OFFSET,
 759        "ON": TokenType.ON,
 760        "OR": TokenType.OR,
 761        "XOR": TokenType.XOR,
 762        "ORDER BY": TokenType.ORDER_BY,
 763        "ORDINALITY": TokenType.ORDINALITY,
 764        "OUTER": TokenType.OUTER,
 765        "OVER": TokenType.OVER,
 766        "OVERLAPS": TokenType.OVERLAPS,
 767        "OVERWRITE": TokenType.OVERWRITE,
 768        "PARTITION": TokenType.PARTITION,
 769        "PARTITION BY": TokenType.PARTITION_BY,
 770        "PARTITIONED BY": TokenType.PARTITION_BY,
 771        "PARTITIONED_BY": TokenType.PARTITION_BY,
 772        "PERCENT": TokenType.PERCENT,
 773        "PIVOT": TokenType.PIVOT,
 774        "PRAGMA": TokenType.PRAGMA,
 775        "PRIMARY KEY": TokenType.PRIMARY_KEY,
 776        "PROCEDURE": TokenType.PROCEDURE,
 777        "QUALIFY": TokenType.QUALIFY,
 778        "RANGE": TokenType.RANGE,
 779        "RECURSIVE": TokenType.RECURSIVE,
 780        "REGEXP": TokenType.RLIKE,
 781        "RENAME": TokenType.RENAME,
 782        "REPLACE": TokenType.REPLACE,
 783        "RETURNING": TokenType.RETURNING,
 784        "REFERENCES": TokenType.REFERENCES,
 785        "RIGHT": TokenType.RIGHT,
 786        "RLIKE": TokenType.RLIKE,
 787        "ROLLBACK": TokenType.ROLLBACK,
 788        "ROLLUP": TokenType.ROLLUP,
 789        "ROW": TokenType.ROW,
 790        "ROWS": TokenType.ROWS,
 791        "SCHEMA": TokenType.SCHEMA,
 792        "SELECT": TokenType.SELECT,
 793        "SEMI": TokenType.SEMI,
 794        "SET": TokenType.SET,
 795        "SETTINGS": TokenType.SETTINGS,
 796        "SHOW": TokenType.SHOW,
 797        "SIMILAR TO": TokenType.SIMILAR_TO,
 798        "SOME": TokenType.SOME,
 799        "SORT BY": TokenType.SORT_BY,
 800        "START WITH": TokenType.START_WITH,
 801        "STRAIGHT_JOIN": TokenType.STRAIGHT_JOIN,
 802        "TABLE": TokenType.TABLE,
 803        "TABLESAMPLE": TokenType.TABLE_SAMPLE,
 804        "TEMP": TokenType.TEMPORARY,
 805        "TEMPORARY": TokenType.TEMPORARY,
 806        "THEN": TokenType.THEN,
 807        "TRUE": TokenType.TRUE,
 808        "TRUNCATE": TokenType.TRUNCATE,
 809        "UNION": TokenType.UNION,
 810        "UNKNOWN": TokenType.UNKNOWN,
 811        "UNNEST": TokenType.UNNEST,
 812        "UNPIVOT": TokenType.UNPIVOT,
 813        "UPDATE": TokenType.UPDATE,
 814        "USE": TokenType.USE,
 815        "USING": TokenType.USING,
 816        "UUID": TokenType.UUID,
 817        "VALUES": TokenType.VALUES,
 818        "VIEW": TokenType.VIEW,
 819        "VOLATILE": TokenType.VOLATILE,
 820        "WHEN": TokenType.WHEN,
 821        "WHERE": TokenType.WHERE,
 822        "WINDOW": TokenType.WINDOW,
 823        "WITH": TokenType.WITH,
 824        "APPLY": TokenType.APPLY,
 825        "ARRAY": TokenType.ARRAY,
 826        "BIT": TokenType.BIT,
 827        "BOOL": TokenType.BOOLEAN,
 828        "BOOLEAN": TokenType.BOOLEAN,
 829        "BYTE": TokenType.TINYINT,
 830        "MEDIUMINT": TokenType.MEDIUMINT,
 831        "INT1": TokenType.TINYINT,
 832        "TINYINT": TokenType.TINYINT,
 833        "INT16": TokenType.SMALLINT,
 834        "SHORT": TokenType.SMALLINT,
 835        "SMALLINT": TokenType.SMALLINT,
 836        "INT128": TokenType.INT128,
 837        "HUGEINT": TokenType.INT128,
 838        "UHUGEINT": TokenType.UINT128,
 839        "INT2": TokenType.SMALLINT,
 840        "INTEGER": TokenType.INT,
 841        "INT": TokenType.INT,
 842        "INT4": TokenType.INT,
 843        "INT32": TokenType.INT,
 844        "INT64": TokenType.BIGINT,
 845        "LONG": TokenType.BIGINT,
 846        "BIGINT": TokenType.BIGINT,
 847        "INT8": TokenType.TINYINT,
 848        "UINT": TokenType.UINT,
 849        "DEC": TokenType.DECIMAL,
 850        "DECIMAL": TokenType.DECIMAL,
 851        "DECIMAL32": TokenType.DECIMAL32,
 852        "DECIMAL64": TokenType.DECIMAL64,
 853        "DECIMAL128": TokenType.DECIMAL128,
 854        "DECIMAL256": TokenType.DECIMAL256,
 855        "BIGDECIMAL": TokenType.BIGDECIMAL,
 856        "BIGNUMERIC": TokenType.BIGDECIMAL,
 857        "LIST": TokenType.LIST,
 858        "MAP": TokenType.MAP,
 859        "NULLABLE": TokenType.NULLABLE,
 860        "NUMBER": TokenType.DECIMAL,
 861        "NUMERIC": TokenType.DECIMAL,
 862        "FIXED": TokenType.DECIMAL,
 863        "REAL": TokenType.FLOAT,
 864        "FLOAT": TokenType.FLOAT,
 865        "FLOAT4": TokenType.FLOAT,
 866        "FLOAT8": TokenType.DOUBLE,
 867        "DOUBLE": TokenType.DOUBLE,
 868        "DOUBLE PRECISION": TokenType.DOUBLE,
 869        "JSON": TokenType.JSON,
 870        "JSONB": TokenType.JSONB,
 871        "CHAR": TokenType.CHAR,
 872        "CHARACTER": TokenType.CHAR,
 873        "NCHAR": TokenType.NCHAR,
 874        "VARCHAR": TokenType.VARCHAR,
 875        "VARCHAR2": TokenType.VARCHAR,
 876        "NVARCHAR": TokenType.NVARCHAR,
 877        "NVARCHAR2": TokenType.NVARCHAR,
 878        "BPCHAR": TokenType.BPCHAR,
 879        "STR": TokenType.TEXT,
 880        "STRING": TokenType.TEXT,
 881        "TEXT": TokenType.TEXT,
 882        "LONGTEXT": TokenType.LONGTEXT,
 883        "MEDIUMTEXT": TokenType.MEDIUMTEXT,
 884        "TINYTEXT": TokenType.TINYTEXT,
 885        "CLOB": TokenType.TEXT,
 886        "LONGVARCHAR": TokenType.TEXT,
 887        "BINARY": TokenType.BINARY,
 888        "BLOB": TokenType.VARBINARY,
 889        "LONGBLOB": TokenType.LONGBLOB,
 890        "MEDIUMBLOB": TokenType.MEDIUMBLOB,
 891        "TINYBLOB": TokenType.TINYBLOB,
 892        "BYTEA": TokenType.VARBINARY,
 893        "VARBINARY": TokenType.VARBINARY,
 894        "TIME": TokenType.TIME,
 895        "TIMETZ": TokenType.TIMETZ,
 896        "TIMESTAMP": TokenType.TIMESTAMP,
 897        "TIMESTAMPTZ": TokenType.TIMESTAMPTZ,
 898        "TIMESTAMPLTZ": TokenType.TIMESTAMPLTZ,
 899        "TIMESTAMP_LTZ": TokenType.TIMESTAMPLTZ,
 900        "TIMESTAMPNTZ": TokenType.TIMESTAMPNTZ,
 901        "TIMESTAMP_NTZ": TokenType.TIMESTAMPNTZ,
 902        "DATE": TokenType.DATE,
 903        "DATETIME": TokenType.DATETIME,
 904        "INT4RANGE": TokenType.INT4RANGE,
 905        "INT4MULTIRANGE": TokenType.INT4MULTIRANGE,
 906        "INT8RANGE": TokenType.INT8RANGE,
 907        "INT8MULTIRANGE": TokenType.INT8MULTIRANGE,
 908        "NUMRANGE": TokenType.NUMRANGE,
 909        "NUMMULTIRANGE": TokenType.NUMMULTIRANGE,
 910        "TSRANGE": TokenType.TSRANGE,
 911        "TSMULTIRANGE": TokenType.TSMULTIRANGE,
 912        "TSTZRANGE": TokenType.TSTZRANGE,
 913        "TSTZMULTIRANGE": TokenType.TSTZMULTIRANGE,
 914        "DATERANGE": TokenType.DATERANGE,
 915        "DATEMULTIRANGE": TokenType.DATEMULTIRANGE,
 916        "UNIQUE": TokenType.UNIQUE,
 917        "VECTOR": TokenType.VECTOR,
 918        "STRUCT": TokenType.STRUCT,
 919        "SEQUENCE": TokenType.SEQUENCE,
 920        "VARIANT": TokenType.VARIANT,
 921        "ALTER": TokenType.ALTER,
 922        "ANALYZE": TokenType.COMMAND,
 923        "CALL": TokenType.COMMAND,
 924        "COMMENT": TokenType.COMMENT,
 925        "EXPLAIN": TokenType.COMMAND,
 926        "GRANT": TokenType.GRANT,
 927        "OPTIMIZE": TokenType.COMMAND,
 928        "PREPARE": TokenType.COMMAND,
 929        "VACUUM": TokenType.COMMAND,
 930        "USER-DEFINED": TokenType.USERDEFINED,
 931        "FOR VERSION": TokenType.VERSION_SNAPSHOT,
 932        "FOR TIMESTAMP": TokenType.TIMESTAMP_SNAPSHOT,
 933    }
 934
 935    WHITE_SPACE: t.Dict[t.Optional[str], TokenType] = {
 936        " ": TokenType.SPACE,
 937        "\t": TokenType.SPACE,
 938        "\n": TokenType.BREAK,
 939        "\r": TokenType.BREAK,
 940    }
 941
 942    COMMANDS = {
 943        TokenType.COMMAND,
 944        TokenType.EXECUTE,
 945        TokenType.FETCH,
 946        TokenType.SHOW,
 947        TokenType.RENAME,
 948    }
 949
 950    COMMAND_PREFIX_TOKENS = {TokenType.SEMICOLON, TokenType.BEGIN}
 951
 952    # Handle numeric literals like in hive (3L = BIGINT)
 953    NUMERIC_LITERALS: t.Dict[str, str] = {}
 954
 955    COMMENTS = ["--", ("/*", "*/")]
 956
 957    __slots__ = (
 958        "sql",
 959        "size",
 960        "tokens",
 961        "dialect",
 962        "_start",
 963        "_current",
 964        "_line",
 965        "_col",
 966        "_comments",
 967        "_char",
 968        "_end",
 969        "_peek",
 970        "_prev_token_line",
 971        "_rs_dialect_settings",
 972    )
 973
 974    def __init__(self, dialect: DialectType = None) -> None:
 975        from sqlglot.dialects import Dialect
 976
 977        self.dialect = Dialect.get_or_raise(dialect)
 978
 979        if USE_RS_TOKENIZER:
 980            self._rs_dialect_settings = RsTokenizerDialectSettings(
 981                unescaped_sequences=self.dialect.UNESCAPED_SEQUENCES,
 982                identifiers_can_start_with_digit=self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT,
 983            )
 984
 985        self.reset()
 986
 987    def reset(self) -> None:
 988        self.sql = ""
 989        self.size = 0
 990        self.tokens: t.List[Token] = []
 991        self._start = 0
 992        self._current = 0
 993        self._line = 1
 994        self._col = 0
 995        self._comments: t.List[str] = []
 996
 997        self._char = ""
 998        self._end = False
 999        self._peek = ""
1000        self._prev_token_line = -1
1001
1002    def tokenize(self, sql: str) -> t.List[Token]:
1003        """Returns a list of tokens corresponding to the SQL string `sql`."""
1004        if USE_RS_TOKENIZER:
1005            return self.tokenize_rs(sql)
1006
1007        self.reset()
1008        self.sql = sql
1009        self.size = len(sql)
1010
1011        try:
1012            self._scan()
1013        except Exception as e:
1014            start = max(self._current - 50, 0)
1015            end = min(self._current + 50, self.size - 1)
1016            context = self.sql[start:end]
1017            raise TokenError(f"Error tokenizing '{context}'") from e
1018
1019        return self.tokens
1020
1021    def _scan(self, until: t.Optional[t.Callable] = None) -> None:
1022        while self.size and not self._end:
1023            current = self._current
1024
1025            # Skip spaces here rather than iteratively calling advance() for performance reasons
1026            while current < self.size:
1027                char = self.sql[current]
1028
1029                if char.isspace() and (char == " " or char == "\t"):
1030                    current += 1
1031                else:
1032                    break
1033
1034            offset = current - self._current if current > self._current else 1
1035
1036            self._start = current
1037            self._advance(offset)
1038
1039            if not self._char.isspace():
1040                if self._char.isdigit():
1041                    self._scan_number()
1042                elif self._char in self._IDENTIFIERS:
1043                    self._scan_identifier(self._IDENTIFIERS[self._char])
1044                else:
1045                    self._scan_keywords()
1046
1047            if until and until():
1048                break
1049
1050        if self.tokens and self._comments:
1051            self.tokens[-1].comments.extend(self._comments)
1052
1053    def _chars(self, size: int) -> str:
1054        if size == 1:
1055            return self._char
1056
1057        start = self._current - 1
1058        end = start + size
1059
1060        return self.sql[start:end] if end <= self.size else ""
1061
1062    def _advance(self, i: int = 1, alnum: bool = False) -> None:
1063        if self.WHITE_SPACE.get(self._char) is TokenType.BREAK:
1064            # Ensures we don't count an extra line if we get a \r\n line break sequence
1065            if not (self._char == "\r" and self._peek == "\n"):
1066                self._col = 1
1067                self._line += 1
1068        else:
1069            self._col += i
1070
1071        self._current += i
1072        self._end = self._current >= self.size
1073        self._char = self.sql[self._current - 1]
1074        self._peek = "" if self._end else self.sql[self._current]
1075
1076        if alnum and self._char.isalnum():
1077            # Here we use local variables instead of attributes for better performance
1078            _col = self._col
1079            _current = self._current
1080            _end = self._end
1081            _peek = self._peek
1082
1083            while _peek.isalnum():
1084                _col += 1
1085                _current += 1
1086                _end = _current >= self.size
1087                _peek = "" if _end else self.sql[_current]
1088
1089            self._col = _col
1090            self._current = _current
1091            self._end = _end
1092            self._peek = _peek
1093            self._char = self.sql[_current - 1]
1094
1095    @property
1096    def _text(self) -> str:
1097        return self.sql[self._start : self._current]
1098
1099    def _add(self, token_type: TokenType, text: t.Optional[str] = None) -> None:
1100        self._prev_token_line = self._line
1101
1102        if self._comments and token_type == TokenType.SEMICOLON and self.tokens:
1103            self.tokens[-1].comments.extend(self._comments)
1104            self._comments = []
1105
1106        self.tokens.append(
1107            Token(
1108                token_type,
1109                text=self._text if text is None else text,
1110                line=self._line,
1111                col=self._col,
1112                start=self._start,
1113                end=self._current - 1,
1114                comments=self._comments,
1115            )
1116        )
1117        self._comments = []
1118
1119        # If we have either a semicolon or a begin token before the command's token, we'll parse
1120        # whatever follows the command's token as a string
1121        if (
1122            token_type in self.COMMANDS
1123            and self._peek != ";"
1124            and (len(self.tokens) == 1 or self.tokens[-2].token_type in self.COMMAND_PREFIX_TOKENS)
1125        ):
1126            start = self._current
1127            tokens = len(self.tokens)
1128            self._scan(lambda: self._peek == ";")
1129            self.tokens = self.tokens[:tokens]
1130            text = self.sql[start : self._current].strip()
1131            if text:
1132                self._add(TokenType.STRING, text)
1133
1134    def _scan_keywords(self) -> None:
1135        size = 0
1136        word = None
1137        chars = self._text
1138        char = chars
1139        prev_space = False
1140        skip = False
1141        trie = self._KEYWORD_TRIE
1142        single_token = char in self.SINGLE_TOKENS
1143
1144        while chars:
1145            if skip:
1146                result = TrieResult.PREFIX
1147            else:
1148                result, trie = in_trie(trie, char.upper())
1149
1150            if result == TrieResult.FAILED:
1151                break
1152            if result == TrieResult.EXISTS:
1153                word = chars
1154
1155            end = self._current + size
1156            size += 1
1157
1158            if end < self.size:
1159                char = self.sql[end]
1160                single_token = single_token or char in self.SINGLE_TOKENS
1161                is_space = char.isspace()
1162
1163                if not is_space or not prev_space:
1164                    if is_space:
1165                        char = " "
1166                    chars += char
1167                    prev_space = is_space
1168                    skip = False
1169                else:
1170                    skip = True
1171            else:
1172                char = ""
1173                break
1174
1175        if word:
1176            if self._scan_string(word):
1177                return
1178            if self._scan_comment(word):
1179                return
1180            if prev_space or single_token or not char:
1181                self._advance(size - 1)
1182                word = word.upper()
1183                self._add(self.KEYWORDS[word], text=word)
1184                return
1185
1186        if self._char in self.SINGLE_TOKENS:
1187            self._add(self.SINGLE_TOKENS[self._char], text=self._char)
1188            return
1189
1190        self._scan_var()
1191
1192    def _scan_comment(self, comment_start: str) -> bool:
1193        if comment_start not in self._COMMENTS:
1194            return False
1195
1196        comment_start_line = self._line
1197        comment_start_size = len(comment_start)
1198        comment_end = self._COMMENTS[comment_start]
1199
1200        if comment_end:
1201            # Skip the comment's start delimiter
1202            self._advance(comment_start_size)
1203
1204            comment_count = 1
1205            comment_end_size = len(comment_end)
1206
1207            while not self._end:
1208                if self._chars(comment_end_size) == comment_end:
1209                    comment_count -= 1
1210                    if not comment_count:
1211                        break
1212
1213                self._advance(alnum=True)
1214
1215                # Nested comments are allowed by some dialects, e.g. databricks, duckdb, postgres
1216                if (
1217                    self.NESTED_COMMENTS
1218                    and not self._end
1219                    and self._chars(comment_end_size) == comment_start
1220                ):
1221                    self._advance(comment_start_size)
1222                    comment_count += 1
1223
1224            self._comments.append(self._text[comment_start_size : -comment_end_size + 1])
1225            self._advance(comment_end_size - 1)
1226        else:
1227            while not self._end and self.WHITE_SPACE.get(self._peek) is not TokenType.BREAK:
1228                self._advance(alnum=True)
1229            self._comments.append(self._text[comment_start_size:])
1230
1231        # Leading comment is attached to the succeeding token, whilst trailing comment to the preceding.
1232        # Multiple consecutive comments are preserved by appending them to the current comments list.
1233        if comment_start_line == self._prev_token_line:
1234            self.tokens[-1].comments.extend(self._comments)
1235            self._comments = []
1236            self._prev_token_line = self._line
1237
1238        return True
1239
1240    def _scan_number(self) -> None:
1241        if self._char == "0":
1242            peek = self._peek.upper()
1243            if peek == "B":
1244                return self._scan_bits() if self.BIT_STRINGS else self._add(TokenType.NUMBER)
1245            elif peek == "X":
1246                return self._scan_hex() if self.HEX_STRINGS else self._add(TokenType.NUMBER)
1247
1248        decimal = False
1249        scientific = 0
1250
1251        while True:
1252            if self._peek.isdigit():
1253                self._advance()
1254            elif self._peek == "." and not decimal:
1255                if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER:
1256                    return self._add(TokenType.NUMBER)
1257                decimal = True
1258                self._advance()
1259            elif self._peek in ("-", "+") and scientific == 1:
1260                scientific += 1
1261                self._advance()
1262            elif self._peek.upper() == "E" and not scientific:
1263                scientific += 1
1264                self._advance()
1265            elif self._peek.isidentifier():
1266                number_text = self._text
1267                literal = ""
1268
1269                while self._peek.strip() and self._peek not in self.SINGLE_TOKENS:
1270                    literal += self._peek
1271                    self._advance()
1272
1273                token_type = self.KEYWORDS.get(self.NUMERIC_LITERALS.get(literal.upper(), ""))
1274
1275                if token_type:
1276                    self._add(TokenType.NUMBER, number_text)
1277                    self._add(TokenType.DCOLON, "::")
1278                    return self._add(token_type, literal)
1279                elif self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT:
1280                    return self._add(TokenType.VAR)
1281
1282                self._advance(-len(literal))
1283                return self._add(TokenType.NUMBER, number_text)
1284            else:
1285                return self._add(TokenType.NUMBER)
1286
1287    def _scan_bits(self) -> None:
1288        self._advance()
1289        value = self._extract_value()
1290        try:
1291            # If `value` can't be converted to a binary, fallback to tokenizing it as an identifier
1292            int(value, 2)
1293            self._add(TokenType.BIT_STRING, value[2:])  # Drop the 0b
1294        except ValueError:
1295            self._add(TokenType.IDENTIFIER)
1296
1297    def _scan_hex(self) -> None:
1298        self._advance()
1299        value = self._extract_value()
1300        try:
1301            # If `value` can't be converted to a hex, fallback to tokenizing it as an identifier
1302            int(value, 16)
1303            self._add(TokenType.HEX_STRING, value[2:])  # Drop the 0x
1304        except ValueError:
1305            self._add(TokenType.IDENTIFIER)
1306
1307    def _extract_value(self) -> str:
1308        while True:
1309            char = self._peek.strip()
1310            if char and char not in self.SINGLE_TOKENS:
1311                self._advance(alnum=True)
1312            else:
1313                break
1314
1315        return self._text
1316
1317    def _scan_string(self, start: str) -> bool:
1318        base = None
1319        token_type = TokenType.STRING
1320
1321        if start in self._QUOTES:
1322            end = self._QUOTES[start]
1323        elif start in self._FORMAT_STRINGS:
1324            end, token_type = self._FORMAT_STRINGS[start]
1325
1326            if token_type == TokenType.HEX_STRING:
1327                base = 16
1328            elif token_type == TokenType.BIT_STRING:
1329                base = 2
1330            elif token_type == TokenType.HEREDOC_STRING:
1331                self._advance()
1332
1333                if self._char == end:
1334                    tag = ""
1335                else:
1336                    tag = self._extract_string(
1337                        end,
1338                        raw_string=True,
1339                        raise_unmatched=not self.HEREDOC_TAG_IS_IDENTIFIER,
1340                    )
1341
1342                if tag and self.HEREDOC_TAG_IS_IDENTIFIER and (self._end or not tag.isidentifier()):
1343                    if not self._end:
1344                        self._advance(-1)
1345
1346                    self._advance(-len(tag))
1347                    self._add(self.HEREDOC_STRING_ALTERNATIVE)
1348                    return True
1349
1350                end = f"{start}{tag}{end}"
1351        else:
1352            return False
1353
1354        self._advance(len(start))
1355        text = self._extract_string(end, raw_string=token_type == TokenType.RAW_STRING)
1356
1357        if base:
1358            try:
1359                int(text, base)
1360            except Exception:
1361                raise TokenError(
1362                    f"Numeric string contains invalid characters from {self._line}:{self._start}"
1363                )
1364
1365        self._add(token_type, text)
1366        return True
1367
1368    def _scan_identifier(self, identifier_end: str) -> None:
1369        self._advance()
1370        text = self._extract_string(
1371            identifier_end, escapes=self._IDENTIFIER_ESCAPES | {identifier_end}
1372        )
1373        self._add(TokenType.IDENTIFIER, text)
1374
1375    def _scan_var(self) -> None:
1376        while True:
1377            char = self._peek.strip()
1378            if char and (char in self.VAR_SINGLE_TOKENS or char not in self.SINGLE_TOKENS):
1379                self._advance(alnum=True)
1380            else:
1381                break
1382
1383        self._add(
1384            TokenType.VAR
1385            if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER
1386            else self.KEYWORDS.get(self._text.upper(), TokenType.VAR)
1387        )
1388
1389    def _extract_string(
1390        self,
1391        delimiter: str,
1392        escapes: t.Optional[t.Set[str]] = None,
1393        raw_string: bool = False,
1394        raise_unmatched: bool = True,
1395    ) -> str:
1396        text = ""
1397        delim_size = len(delimiter)
1398        escapes = self._STRING_ESCAPES if escapes is None else escapes
1399
1400        while True:
1401            if (
1402                not raw_string
1403                and self.dialect.UNESCAPED_SEQUENCES
1404                and self._peek
1405                and self._char in self.STRING_ESCAPES
1406            ):
1407                unescaped_sequence = self.dialect.UNESCAPED_SEQUENCES.get(self._char + self._peek)
1408                if unescaped_sequence:
1409                    self._advance(2)
1410                    text += unescaped_sequence
1411                    continue
1412            if (
1413                (self.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS or not raw_string)
1414                and self._char in escapes
1415                and (self._peek == delimiter or self._peek in escapes)
1416                and (self._char not in self._QUOTES or self._char == self._peek)
1417            ):
1418                if self._peek == delimiter:
1419                    text += self._peek
1420                else:
1421                    text += self._char + self._peek
1422
1423                if self._current + 1 < self.size:
1424                    self._advance(2)
1425                else:
1426                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._current}")
1427            else:
1428                if self._chars(delim_size) == delimiter:
1429                    if delim_size > 1:
1430                        self._advance(delim_size - 1)
1431                    break
1432
1433                if self._end:
1434                    if not raise_unmatched:
1435                        return text + self._char
1436
1437                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._start}")
1438
1439                current = self._current - 1
1440                self._advance(alnum=True)
1441                text += self.sql[current : self._current - 1]
1442
1443        return text
1444
1445    def tokenize_rs(self, sql: str) -> t.List[Token]:
1446        if not self._RS_TOKENIZER:
1447            raise SqlglotError("Rust tokenizer is not available")
1448
1449        try:
1450            tokens = self._RS_TOKENIZER.tokenize(sql, self._rs_dialect_settings)
1451            for token in tokens:
1452                token.token_type = _ALL_TOKEN_TYPES[token.token_type_index]
1453            return tokens
1454        except Exception as e:
1455            raise TokenError(str(e))