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    DOTCOLON = auto()
  41    DCOLON = auto()
  42    DQMARK = auto()
  43    SEMICOLON = auto()
  44    STAR = auto()
  45    BACKSLASH = auto()
  46    SLASH = auto()
  47    LT = auto()
  48    LTE = auto()
  49    GT = auto()
  50    GTE = auto()
  51    NOT = auto()
  52    EQ = auto()
  53    NEQ = auto()
  54    NULLSAFE_EQ = auto()
  55    COLON_EQ = auto()
  56    AND = auto()
  57    OR = auto()
  58    AMP = auto()
  59    DPIPE = auto()
  60    PIPE = auto()
  61    PIPE_SLASH = auto()
  62    DPIPE_SLASH = auto()
  63    CARET = auto()
  64    CARET_AT = auto()
  65    TILDA = auto()
  66    ARROW = auto()
  67    DARROW = auto()
  68    FARROW = auto()
  69    HASH = auto()
  70    HASH_ARROW = auto()
  71    DHASH_ARROW = auto()
  72    LR_ARROW = auto()
  73    DAT = auto()
  74    LT_AT = auto()
  75    AT_GT = auto()
  76    DOLLAR = auto()
  77    PARAMETER = auto()
  78    SESSION_PARAMETER = auto()
  79    DAMP = auto()
  80    XOR = auto()
  81    DSTAR = auto()
  82
  83    URI_START = auto()
  84
  85    BLOCK_START = auto()
  86    BLOCK_END = auto()
  87
  88    SPACE = auto()
  89    BREAK = auto()
  90
  91    STRING = auto()
  92    NUMBER = auto()
  93    IDENTIFIER = auto()
  94    DATABASE = auto()
  95    COLUMN = auto()
  96    COLUMN_DEF = auto()
  97    SCHEMA = auto()
  98    TABLE = auto()
  99    WAREHOUSE = auto()
 100    STAGE = auto()
 101    STREAMLIT = auto()
 102    VAR = auto()
 103    BIT_STRING = auto()
 104    HEX_STRING = auto()
 105    BYTE_STRING = auto()
 106    NATIONAL_STRING = auto()
 107    RAW_STRING = auto()
 108    HEREDOC_STRING = auto()
 109    UNICODE_STRING = auto()
 110
 111    # types
 112    BIT = auto()
 113    BOOLEAN = auto()
 114    TINYINT = auto()
 115    UTINYINT = auto()
 116    SMALLINT = auto()
 117    USMALLINT = auto()
 118    MEDIUMINT = auto()
 119    UMEDIUMINT = auto()
 120    INT = auto()
 121    UINT = auto()
 122    BIGINT = auto()
 123    UBIGINT = auto()
 124    INT128 = auto()
 125    UINT128 = auto()
 126    INT256 = auto()
 127    UINT256 = auto()
 128    FLOAT = auto()
 129    DOUBLE = auto()
 130    UDOUBLE = auto()
 131    DECIMAL = auto()
 132    DECIMAL32 = auto()
 133    DECIMAL64 = auto()
 134    DECIMAL128 = auto()
 135    DECIMAL256 = auto()
 136    UDECIMAL = auto()
 137    BIGDECIMAL = auto()
 138    CHAR = auto()
 139    NCHAR = auto()
 140    VARCHAR = auto()
 141    NVARCHAR = auto()
 142    BPCHAR = auto()
 143    TEXT = auto()
 144    MEDIUMTEXT = auto()
 145    LONGTEXT = auto()
 146    BLOB = auto()
 147    MEDIUMBLOB = auto()
 148    LONGBLOB = auto()
 149    TINYBLOB = auto()
 150    TINYTEXT = auto()
 151    NAME = auto()
 152    BINARY = auto()
 153    VARBINARY = auto()
 154    JSON = auto()
 155    JSONB = auto()
 156    TIME = auto()
 157    TIMETZ = auto()
 158    TIMESTAMP = auto()
 159    TIMESTAMPTZ = auto()
 160    TIMESTAMPLTZ = auto()
 161    TIMESTAMPNTZ = auto()
 162    TIMESTAMP_S = auto()
 163    TIMESTAMP_MS = auto()
 164    TIMESTAMP_NS = auto()
 165    DATETIME = auto()
 166    DATETIME2 = auto()
 167    DATETIME64 = auto()
 168    SMALLDATETIME = auto()
 169    DATE = auto()
 170    DATE32 = auto()
 171    INT4RANGE = auto()
 172    INT4MULTIRANGE = auto()
 173    INT8RANGE = auto()
 174    INT8MULTIRANGE = auto()
 175    NUMRANGE = auto()
 176    NUMMULTIRANGE = auto()
 177    TSRANGE = auto()
 178    TSMULTIRANGE = auto()
 179    TSTZRANGE = auto()
 180    TSTZMULTIRANGE = auto()
 181    DATERANGE = auto()
 182    DATEMULTIRANGE = auto()
 183    UUID = auto()
 184    GEOGRAPHY = auto()
 185    NULLABLE = auto()
 186    GEOMETRY = auto()
 187    POINT = auto()
 188    RING = auto()
 189    LINESTRING = auto()
 190    MULTILINESTRING = auto()
 191    POLYGON = auto()
 192    MULTIPOLYGON = auto()
 193    HLLSKETCH = auto()
 194    HSTORE = auto()
 195    SUPER = auto()
 196    SERIAL = auto()
 197    SMALLSERIAL = auto()
 198    BIGSERIAL = auto()
 199    XML = auto()
 200    YEAR = auto()
 201    USERDEFINED = auto()
 202    MONEY = auto()
 203    SMALLMONEY = auto()
 204    ROWVERSION = auto()
 205    IMAGE = auto()
 206    VARIANT = auto()
 207    OBJECT = auto()
 208    INET = auto()
 209    IPADDRESS = auto()
 210    IPPREFIX = auto()
 211    IPV4 = auto()
 212    IPV6 = auto()
 213    ENUM = auto()
 214    ENUM8 = auto()
 215    ENUM16 = auto()
 216    FIXEDSTRING = auto()
 217    LOWCARDINALITY = auto()
 218    NESTED = auto()
 219    AGGREGATEFUNCTION = auto()
 220    SIMPLEAGGREGATEFUNCTION = auto()
 221    TDIGEST = auto()
 222    UNKNOWN = auto()
 223    VECTOR = auto()
 224    DYNAMIC = auto()
 225    VOID = auto()
 226
 227    # keywords
 228    ALIAS = auto()
 229    ALTER = auto()
 230    ALWAYS = auto()
 231    ALL = auto()
 232    ANTI = auto()
 233    ANY = auto()
 234    APPLY = auto()
 235    ARRAY = auto()
 236    ASC = auto()
 237    ASOF = auto()
 238    ATTACH = auto()
 239    AUTO_INCREMENT = auto()
 240    BEGIN = auto()
 241    BETWEEN = auto()
 242    BULK_COLLECT_INTO = auto()
 243    CACHE = auto()
 244    CASE = auto()
 245    CHARACTER_SET = auto()
 246    CLUSTER_BY = auto()
 247    COLLATE = auto()
 248    COMMAND = auto()
 249    COMMENT = auto()
 250    COMMIT = auto()
 251    CONNECT_BY = auto()
 252    CONSTRAINT = auto()
 253    COPY = auto()
 254    CREATE = auto()
 255    CROSS = auto()
 256    CUBE = auto()
 257    CURRENT_DATE = auto()
 258    CURRENT_DATETIME = auto()
 259    CURRENT_SCHEMA = auto()
 260    CURRENT_TIME = auto()
 261    CURRENT_TIMESTAMP = auto()
 262    CURRENT_USER = auto()
 263    DECLARE = auto()
 264    DEFAULT = auto()
 265    DELETE = auto()
 266    DESC = auto()
 267    DESCRIBE = auto()
 268    DETACH = auto()
 269    DICTIONARY = auto()
 270    DISTINCT = auto()
 271    DISTRIBUTE_BY = auto()
 272    DIV = auto()
 273    DROP = auto()
 274    ELSE = auto()
 275    END = auto()
 276    ESCAPE = auto()
 277    EXCEPT = auto()
 278    EXECUTE = auto()
 279    EXISTS = auto()
 280    FALSE = auto()
 281    FETCH = auto()
 282    FILE_FORMAT = auto()
 283    FILTER = auto()
 284    FINAL = auto()
 285    FIRST = auto()
 286    FOR = auto()
 287    FORCE = auto()
 288    FOREIGN_KEY = auto()
 289    FORMAT = auto()
 290    FROM = auto()
 291    FULL = auto()
 292    FUNCTION = auto()
 293    GLOB = auto()
 294    GLOBAL = auto()
 295    GRANT = auto()
 296    GROUP_BY = auto()
 297    GROUPING_SETS = auto()
 298    HAVING = auto()
 299    HINT = auto()
 300    IGNORE = auto()
 301    ILIKE = auto()
 302    ILIKE_ANY = auto()
 303    IN = auto()
 304    INDEX = auto()
 305    INNER = auto()
 306    INSERT = auto()
 307    INTERSECT = auto()
 308    INTERVAL = auto()
 309    INTO = auto()
 310    INTRODUCER = auto()
 311    IRLIKE = auto()
 312    IS = auto()
 313    ISNULL = auto()
 314    JOIN = auto()
 315    JOIN_MARKER = auto()
 316    KEEP = auto()
 317    KEY = auto()
 318    KILL = auto()
 319    LANGUAGE = auto()
 320    LATERAL = auto()
 321    LEFT = auto()
 322    LIKE = auto()
 323    LIKE_ANY = auto()
 324    LIMIT = auto()
 325    LIST = auto()
 326    LOAD = auto()
 327    LOCK = auto()
 328    MAP = auto()
 329    MATCH_CONDITION = auto()
 330    MATCH_RECOGNIZE = auto()
 331    MEMBER_OF = auto()
 332    MERGE = auto()
 333    MOD = auto()
 334    MODEL = auto()
 335    NATURAL = auto()
 336    NEXT = auto()
 337    NOTHING = auto()
 338    NOTNULL = auto()
 339    NULL = auto()
 340    OBJECT_IDENTIFIER = auto()
 341    OFFSET = auto()
 342    ON = auto()
 343    ONLY = auto()
 344    OPERATOR = auto()
 345    ORDER_BY = auto()
 346    ORDER_SIBLINGS_BY = auto()
 347    ORDERED = auto()
 348    ORDINALITY = auto()
 349    OUTER = auto()
 350    OVER = auto()
 351    OVERLAPS = auto()
 352    OVERWRITE = auto()
 353    PARTITION = auto()
 354    PARTITION_BY = auto()
 355    PERCENT = auto()
 356    PIVOT = auto()
 357    PLACEHOLDER = auto()
 358    POSITIONAL = auto()
 359    PRAGMA = auto()
 360    PREWHERE = auto()
 361    PRIMARY_KEY = auto()
 362    PROCEDURE = auto()
 363    PROPERTIES = auto()
 364    PSEUDO_TYPE = auto()
 365    PUT = auto()
 366    QUALIFY = auto()
 367    QUOTE = auto()
 368    RANGE = auto()
 369    RECURSIVE = auto()
 370    REFRESH = auto()
 371    RENAME = auto()
 372    REPLACE = auto()
 373    RETURNING = auto()
 374    REFERENCES = auto()
 375    RIGHT = auto()
 376    RLIKE = auto()
 377    ROLLBACK = auto()
 378    ROLLUP = auto()
 379    ROW = auto()
 380    ROWS = auto()
 381    SELECT = auto()
 382    SEMI = auto()
 383    SEPARATOR = auto()
 384    SEQUENCE = auto()
 385    SERDE_PROPERTIES = auto()
 386    SET = auto()
 387    SETTINGS = auto()
 388    SHOW = auto()
 389    SIMILAR_TO = auto()
 390    SOME = auto()
 391    SORT_BY = auto()
 392    START_WITH = auto()
 393    STORAGE_INTEGRATION = auto()
 394    STRAIGHT_JOIN = auto()
 395    STRUCT = auto()
 396    SUMMARIZE = auto()
 397    TABLE_SAMPLE = auto()
 398    TAG = auto()
 399    TEMPORARY = auto()
 400    TOP = auto()
 401    THEN = auto()
 402    TRUE = auto()
 403    TRUNCATE = auto()
 404    UNCACHE = auto()
 405    UNION = auto()
 406    UNNEST = auto()
 407    UNPIVOT = auto()
 408    UPDATE = auto()
 409    USE = auto()
 410    USING = auto()
 411    VALUES = auto()
 412    VIEW = auto()
 413    VOLATILE = auto()
 414    WHEN = auto()
 415    WHERE = auto()
 416    WINDOW = auto()
 417    WITH = auto()
 418    UNIQUE = auto()
 419    VERSION_SNAPSHOT = auto()
 420    TIMESTAMP_SNAPSHOT = auto()
 421    OPTION = auto()
 422    SINK = auto()
 423    SOURCE = auto()
 424    ANALYZE = auto()
 425    NAMESPACE = auto()
 426    EXPORT = auto()
 427
 428
 429_ALL_TOKEN_TYPES = list(TokenType)
 430_TOKEN_TYPE_TO_INDEX = {token_type: i for i, token_type in enumerate(_ALL_TOKEN_TYPES)}
 431
 432
 433class Token:
 434    __slots__ = ("token_type", "text", "line", "col", "start", "end", "comments")
 435
 436    @classmethod
 437    def number(cls, number: int) -> Token:
 438        """Returns a NUMBER token with `number` as its text."""
 439        return cls(TokenType.NUMBER, str(number))
 440
 441    @classmethod
 442    def string(cls, string: str) -> Token:
 443        """Returns a STRING token with `string` as its text."""
 444        return cls(TokenType.STRING, string)
 445
 446    @classmethod
 447    def identifier(cls, identifier: str) -> Token:
 448        """Returns an IDENTIFIER token with `identifier` as its text."""
 449        return cls(TokenType.IDENTIFIER, identifier)
 450
 451    @classmethod
 452    def var(cls, var: str) -> Token:
 453        """Returns an VAR token with `var` as its text."""
 454        return cls(TokenType.VAR, var)
 455
 456    def __init__(
 457        self,
 458        token_type: TokenType,
 459        text: str,
 460        line: int = 1,
 461        col: int = 1,
 462        start: int = 0,
 463        end: int = 0,
 464        comments: t.Optional[t.List[str]] = None,
 465    ) -> None:
 466        """Token initializer.
 467
 468        Args:
 469            token_type: The TokenType Enum.
 470            text: The text of the token.
 471            line: The line that the token ends on.
 472            col: The column that the token ends on.
 473            start: The start index of the token.
 474            end: The ending index of the token.
 475            comments: The comments to attach to the token.
 476        """
 477        self.token_type = token_type
 478        self.text = text
 479        self.line = line
 480        self.col = col
 481        self.start = start
 482        self.end = end
 483        self.comments = [] if comments is None else comments
 484
 485    def __repr__(self) -> str:
 486        attributes = ", ".join(f"{k}: {getattr(self, k)}" for k in self.__slots__)
 487        return f"<Token {attributes}>"
 488
 489
 490class _Tokenizer(type):
 491    def __new__(cls, clsname, bases, attrs):
 492        klass = super().__new__(cls, clsname, bases, attrs)
 493
 494        def _convert_quotes(arr: t.List[str | t.Tuple[str, str]]) -> t.Dict[str, str]:
 495            return dict(
 496                (item, item) if isinstance(item, str) else (item[0], item[1]) for item in arr
 497            )
 498
 499        def _quotes_to_format(
 500            token_type: TokenType, arr: t.List[str | t.Tuple[str, str]]
 501        ) -> t.Dict[str, t.Tuple[str, TokenType]]:
 502            return {k: (v, token_type) for k, v in _convert_quotes(arr).items()}
 503
 504        klass._QUOTES = _convert_quotes(klass.QUOTES)
 505        klass._IDENTIFIERS = _convert_quotes(klass.IDENTIFIERS)
 506
 507        klass._FORMAT_STRINGS = {
 508            **{
 509                p + s: (e, TokenType.NATIONAL_STRING)
 510                for s, e in klass._QUOTES.items()
 511                for p in ("n", "N")
 512            },
 513            **_quotes_to_format(TokenType.BIT_STRING, klass.BIT_STRINGS),
 514            **_quotes_to_format(TokenType.BYTE_STRING, klass.BYTE_STRINGS),
 515            **_quotes_to_format(TokenType.HEX_STRING, klass.HEX_STRINGS),
 516            **_quotes_to_format(TokenType.RAW_STRING, klass.RAW_STRINGS),
 517            **_quotes_to_format(TokenType.HEREDOC_STRING, klass.HEREDOC_STRINGS),
 518            **_quotes_to_format(TokenType.UNICODE_STRING, klass.UNICODE_STRINGS),
 519        }
 520
 521        klass._STRING_ESCAPES = set(klass.STRING_ESCAPES)
 522        klass._IDENTIFIER_ESCAPES = set(klass.IDENTIFIER_ESCAPES)
 523        klass._COMMENTS = {
 524            **dict(
 525                (comment, None) if isinstance(comment, str) else (comment[0], comment[1])
 526                for comment in klass.COMMENTS
 527            ),
 528            "{#": "#}",  # Ensure Jinja comments are tokenized correctly in all dialects
 529        }
 530        if klass.HINT_START in klass.KEYWORDS:
 531            klass._COMMENTS[klass.HINT_START] = "*/"
 532
 533        klass._KEYWORD_TRIE = new_trie(
 534            key.upper()
 535            for key in (
 536                *klass.KEYWORDS,
 537                *klass._COMMENTS,
 538                *klass._QUOTES,
 539                *klass._FORMAT_STRINGS,
 540            )
 541            if " " in key or any(single in key for single in klass.SINGLE_TOKENS)
 542        )
 543
 544        if USE_RS_TOKENIZER:
 545            settings = RsTokenizerSettings(
 546                white_space={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.WHITE_SPACE.items()},
 547                single_tokens={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.SINGLE_TOKENS.items()},
 548                keywords={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.KEYWORDS.items()},
 549                numeric_literals=klass.NUMERIC_LITERALS,
 550                identifiers=klass._IDENTIFIERS,
 551                identifier_escapes=klass._IDENTIFIER_ESCAPES,
 552                string_escapes=klass._STRING_ESCAPES,
 553                quotes=klass._QUOTES,
 554                format_strings={
 555                    k: (v1, _TOKEN_TYPE_TO_INDEX[v2])
 556                    for k, (v1, v2) in klass._FORMAT_STRINGS.items()
 557                },
 558                has_bit_strings=bool(klass.BIT_STRINGS),
 559                has_hex_strings=bool(klass.HEX_STRINGS),
 560                comments=klass._COMMENTS,
 561                var_single_tokens=klass.VAR_SINGLE_TOKENS,
 562                commands={_TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMANDS},
 563                command_prefix_tokens={
 564                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMAND_PREFIX_TOKENS
 565                },
 566                heredoc_tag_is_identifier=klass.HEREDOC_TAG_IS_IDENTIFIER,
 567                string_escapes_allowed_in_raw_strings=klass.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS,
 568                nested_comments=klass.NESTED_COMMENTS,
 569                hint_start=klass.HINT_START,
 570                tokens_preceding_hint={
 571                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.TOKENS_PRECEDING_HINT
 572                },
 573            )
 574            token_types = RsTokenTypeSettings(
 575                bit_string=_TOKEN_TYPE_TO_INDEX[TokenType.BIT_STRING],
 576                break_=_TOKEN_TYPE_TO_INDEX[TokenType.BREAK],
 577                dcolon=_TOKEN_TYPE_TO_INDEX[TokenType.DCOLON],
 578                heredoc_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEREDOC_STRING],
 579                raw_string=_TOKEN_TYPE_TO_INDEX[TokenType.RAW_STRING],
 580                hex_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEX_STRING],
 581                identifier=_TOKEN_TYPE_TO_INDEX[TokenType.IDENTIFIER],
 582                number=_TOKEN_TYPE_TO_INDEX[TokenType.NUMBER],
 583                parameter=_TOKEN_TYPE_TO_INDEX[TokenType.PARAMETER],
 584                semicolon=_TOKEN_TYPE_TO_INDEX[TokenType.SEMICOLON],
 585                string=_TOKEN_TYPE_TO_INDEX[TokenType.STRING],
 586                var=_TOKEN_TYPE_TO_INDEX[TokenType.VAR],
 587                heredoc_string_alternative=_TOKEN_TYPE_TO_INDEX[klass.HEREDOC_STRING_ALTERNATIVE],
 588                hint=_TOKEN_TYPE_TO_INDEX[TokenType.HINT],
 589            )
 590            klass._RS_TOKENIZER = RsTokenizer(settings, token_types)
 591        else:
 592            klass._RS_TOKENIZER = None
 593
 594        return klass
 595
 596
 597class Tokenizer(metaclass=_Tokenizer):
 598    SINGLE_TOKENS = {
 599        "(": TokenType.L_PAREN,
 600        ")": TokenType.R_PAREN,
 601        "[": TokenType.L_BRACKET,
 602        "]": TokenType.R_BRACKET,
 603        "{": TokenType.L_BRACE,
 604        "}": TokenType.R_BRACE,
 605        "&": TokenType.AMP,
 606        "^": TokenType.CARET,
 607        ":": TokenType.COLON,
 608        ",": TokenType.COMMA,
 609        ".": TokenType.DOT,
 610        "-": TokenType.DASH,
 611        "=": TokenType.EQ,
 612        ">": TokenType.GT,
 613        "<": TokenType.LT,
 614        "%": TokenType.MOD,
 615        "!": TokenType.NOT,
 616        "|": TokenType.PIPE,
 617        "+": TokenType.PLUS,
 618        ";": TokenType.SEMICOLON,
 619        "/": TokenType.SLASH,
 620        "\\": TokenType.BACKSLASH,
 621        "*": TokenType.STAR,
 622        "~": TokenType.TILDA,
 623        "?": TokenType.PLACEHOLDER,
 624        "@": TokenType.PARAMETER,
 625        "#": TokenType.HASH,
 626        # Used for breaking a var like x'y' but nothing else the token type doesn't matter
 627        "'": TokenType.UNKNOWN,
 628        "`": TokenType.UNKNOWN,
 629        '"': TokenType.UNKNOWN,
 630    }
 631
 632    BIT_STRINGS: t.List[str | t.Tuple[str, str]] = []
 633    BYTE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 634    HEX_STRINGS: t.List[str | t.Tuple[str, str]] = []
 635    RAW_STRINGS: t.List[str | t.Tuple[str, str]] = []
 636    HEREDOC_STRINGS: t.List[str | t.Tuple[str, str]] = []
 637    UNICODE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 638    IDENTIFIERS: t.List[str | t.Tuple[str, str]] = ['"']
 639    QUOTES: t.List[t.Tuple[str, str] | str] = ["'"]
 640    STRING_ESCAPES = ["'"]
 641    VAR_SINGLE_TOKENS: t.Set[str] = set()
 642
 643    # The strings in this list can always be used as escapes, regardless of the surrounding
 644    # identifier delimiters. By default, the closing delimiter is assumed to also act as an
 645    # identifier escape, e.g. if we use double-quotes, then they also act as escapes: "x"""
 646    IDENTIFIER_ESCAPES: t.List[str] = []
 647
 648    # Whether the heredoc tags follow the same lexical rules as unquoted identifiers
 649    HEREDOC_TAG_IS_IDENTIFIER = False
 650
 651    # Token that we'll generate as a fallback if the heredoc prefix doesn't correspond to a heredoc
 652    HEREDOC_STRING_ALTERNATIVE = TokenType.VAR
 653
 654    # Whether string escape characters function as such when placed within raw strings
 655    STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS = True
 656
 657    NESTED_COMMENTS = True
 658
 659    HINT_START = "/*+"
 660
 661    TOKENS_PRECEDING_HINT = {TokenType.SELECT, TokenType.INSERT, TokenType.UPDATE, TokenType.DELETE}
 662
 663    # Autofilled
 664    _COMMENTS: t.Dict[str, str] = {}
 665    _FORMAT_STRINGS: t.Dict[str, t.Tuple[str, TokenType]] = {}
 666    _IDENTIFIERS: t.Dict[str, str] = {}
 667    _IDENTIFIER_ESCAPES: t.Set[str] = set()
 668    _QUOTES: t.Dict[str, str] = {}
 669    _STRING_ESCAPES: t.Set[str] = set()
 670    _KEYWORD_TRIE: t.Dict = {}
 671    _RS_TOKENIZER: t.Optional[t.Any] = None
 672
 673    KEYWORDS: t.Dict[str, TokenType] = {
 674        **{f"{{%{postfix}": TokenType.BLOCK_START for postfix in ("", "+", "-")},
 675        **{f"{prefix}%}}": TokenType.BLOCK_END for prefix in ("", "+", "-")},
 676        **{f"{{{{{postfix}": TokenType.BLOCK_START for postfix in ("+", "-")},
 677        **{f"{prefix}}}}}": TokenType.BLOCK_END for prefix in ("+", "-")},
 678        HINT_START: TokenType.HINT,
 679        "==": TokenType.EQ,
 680        "::": TokenType.DCOLON,
 681        "||": TokenType.DPIPE,
 682        ">=": TokenType.GTE,
 683        "<=": TokenType.LTE,
 684        "<>": TokenType.NEQ,
 685        "!=": TokenType.NEQ,
 686        ":=": TokenType.COLON_EQ,
 687        "<=>": TokenType.NULLSAFE_EQ,
 688        "->": TokenType.ARROW,
 689        "->>": TokenType.DARROW,
 690        "=>": TokenType.FARROW,
 691        "#>": TokenType.HASH_ARROW,
 692        "#>>": TokenType.DHASH_ARROW,
 693        "<->": TokenType.LR_ARROW,
 694        "&&": TokenType.DAMP,
 695        "??": TokenType.DQMARK,
 696        "~~~": TokenType.GLOB,
 697        "~~": TokenType.LIKE,
 698        "~~*": TokenType.ILIKE,
 699        "~*": TokenType.IRLIKE,
 700        "ALL": TokenType.ALL,
 701        "ALWAYS": TokenType.ALWAYS,
 702        "AND": TokenType.AND,
 703        "ANTI": TokenType.ANTI,
 704        "ANY": TokenType.ANY,
 705        "ASC": TokenType.ASC,
 706        "AS": TokenType.ALIAS,
 707        "ASOF": TokenType.ASOF,
 708        "AUTOINCREMENT": TokenType.AUTO_INCREMENT,
 709        "AUTO_INCREMENT": TokenType.AUTO_INCREMENT,
 710        "BEGIN": TokenType.BEGIN,
 711        "BETWEEN": TokenType.BETWEEN,
 712        "CACHE": TokenType.CACHE,
 713        "UNCACHE": TokenType.UNCACHE,
 714        "CASE": TokenType.CASE,
 715        "CHARACTER SET": TokenType.CHARACTER_SET,
 716        "CLUSTER BY": TokenType.CLUSTER_BY,
 717        "COLLATE": TokenType.COLLATE,
 718        "COLUMN": TokenType.COLUMN,
 719        "COMMIT": TokenType.COMMIT,
 720        "CONNECT BY": TokenType.CONNECT_BY,
 721        "CONSTRAINT": TokenType.CONSTRAINT,
 722        "COPY": TokenType.COPY,
 723        "CREATE": TokenType.CREATE,
 724        "CROSS": TokenType.CROSS,
 725        "CUBE": TokenType.CUBE,
 726        "CURRENT_DATE": TokenType.CURRENT_DATE,
 727        "CURRENT_SCHEMA": TokenType.CURRENT_SCHEMA,
 728        "CURRENT_TIME": TokenType.CURRENT_TIME,
 729        "CURRENT_TIMESTAMP": TokenType.CURRENT_TIMESTAMP,
 730        "CURRENT_USER": TokenType.CURRENT_USER,
 731        "DATABASE": TokenType.DATABASE,
 732        "DEFAULT": TokenType.DEFAULT,
 733        "DELETE": TokenType.DELETE,
 734        "DESC": TokenType.DESC,
 735        "DESCRIBE": TokenType.DESCRIBE,
 736        "DISTINCT": TokenType.DISTINCT,
 737        "DISTRIBUTE BY": TokenType.DISTRIBUTE_BY,
 738        "DIV": TokenType.DIV,
 739        "DROP": TokenType.DROP,
 740        "ELSE": TokenType.ELSE,
 741        "END": TokenType.END,
 742        "ENUM": TokenType.ENUM,
 743        "ESCAPE": TokenType.ESCAPE,
 744        "EXCEPT": TokenType.EXCEPT,
 745        "EXECUTE": TokenType.EXECUTE,
 746        "EXISTS": TokenType.EXISTS,
 747        "FALSE": TokenType.FALSE,
 748        "FETCH": TokenType.FETCH,
 749        "FILTER": TokenType.FILTER,
 750        "FIRST": TokenType.FIRST,
 751        "FULL": TokenType.FULL,
 752        "FUNCTION": TokenType.FUNCTION,
 753        "FOR": TokenType.FOR,
 754        "FOREIGN KEY": TokenType.FOREIGN_KEY,
 755        "FORMAT": TokenType.FORMAT,
 756        "FROM": TokenType.FROM,
 757        "GEOGRAPHY": TokenType.GEOGRAPHY,
 758        "GEOMETRY": TokenType.GEOMETRY,
 759        "GLOB": TokenType.GLOB,
 760        "GROUP BY": TokenType.GROUP_BY,
 761        "GROUPING SETS": TokenType.GROUPING_SETS,
 762        "HAVING": TokenType.HAVING,
 763        "ILIKE": TokenType.ILIKE,
 764        "IN": TokenType.IN,
 765        "INDEX": TokenType.INDEX,
 766        "INET": TokenType.INET,
 767        "INNER": TokenType.INNER,
 768        "INSERT": TokenType.INSERT,
 769        "INTERVAL": TokenType.INTERVAL,
 770        "INTERSECT": TokenType.INTERSECT,
 771        "INTO": TokenType.INTO,
 772        "IS": TokenType.IS,
 773        "ISNULL": TokenType.ISNULL,
 774        "JOIN": TokenType.JOIN,
 775        "KEEP": TokenType.KEEP,
 776        "KILL": TokenType.KILL,
 777        "LATERAL": TokenType.LATERAL,
 778        "LEFT": TokenType.LEFT,
 779        "LIKE": TokenType.LIKE,
 780        "LIMIT": TokenType.LIMIT,
 781        "LOAD": TokenType.LOAD,
 782        "LOCK": TokenType.LOCK,
 783        "MERGE": TokenType.MERGE,
 784        "NAMESPACE": TokenType.NAMESPACE,
 785        "NATURAL": TokenType.NATURAL,
 786        "NEXT": TokenType.NEXT,
 787        "NOT": TokenType.NOT,
 788        "NOTNULL": TokenType.NOTNULL,
 789        "NULL": TokenType.NULL,
 790        "OBJECT": TokenType.OBJECT,
 791        "OFFSET": TokenType.OFFSET,
 792        "ON": TokenType.ON,
 793        "OR": TokenType.OR,
 794        "XOR": TokenType.XOR,
 795        "ORDER BY": TokenType.ORDER_BY,
 796        "ORDINALITY": TokenType.ORDINALITY,
 797        "OUTER": TokenType.OUTER,
 798        "OVER": TokenType.OVER,
 799        "OVERLAPS": TokenType.OVERLAPS,
 800        "OVERWRITE": TokenType.OVERWRITE,
 801        "PARTITION": TokenType.PARTITION,
 802        "PARTITION BY": TokenType.PARTITION_BY,
 803        "PARTITIONED BY": TokenType.PARTITION_BY,
 804        "PARTITIONED_BY": TokenType.PARTITION_BY,
 805        "PERCENT": TokenType.PERCENT,
 806        "PIVOT": TokenType.PIVOT,
 807        "PRAGMA": TokenType.PRAGMA,
 808        "PRIMARY KEY": TokenType.PRIMARY_KEY,
 809        "PROCEDURE": TokenType.PROCEDURE,
 810        "QUALIFY": TokenType.QUALIFY,
 811        "RANGE": TokenType.RANGE,
 812        "RECURSIVE": TokenType.RECURSIVE,
 813        "REGEXP": TokenType.RLIKE,
 814        "RENAME": TokenType.RENAME,
 815        "REPLACE": TokenType.REPLACE,
 816        "RETURNING": TokenType.RETURNING,
 817        "REFERENCES": TokenType.REFERENCES,
 818        "RIGHT": TokenType.RIGHT,
 819        "RLIKE": TokenType.RLIKE,
 820        "ROLLBACK": TokenType.ROLLBACK,
 821        "ROLLUP": TokenType.ROLLUP,
 822        "ROW": TokenType.ROW,
 823        "ROWS": TokenType.ROWS,
 824        "SCHEMA": TokenType.SCHEMA,
 825        "SELECT": TokenType.SELECT,
 826        "SEMI": TokenType.SEMI,
 827        "SET": TokenType.SET,
 828        "SETTINGS": TokenType.SETTINGS,
 829        "SHOW": TokenType.SHOW,
 830        "SIMILAR TO": TokenType.SIMILAR_TO,
 831        "SOME": TokenType.SOME,
 832        "SORT BY": TokenType.SORT_BY,
 833        "START WITH": TokenType.START_WITH,
 834        "STRAIGHT_JOIN": TokenType.STRAIGHT_JOIN,
 835        "TABLE": TokenType.TABLE,
 836        "TABLESAMPLE": TokenType.TABLE_SAMPLE,
 837        "TEMP": TokenType.TEMPORARY,
 838        "TEMPORARY": TokenType.TEMPORARY,
 839        "THEN": TokenType.THEN,
 840        "TRUE": TokenType.TRUE,
 841        "TRUNCATE": TokenType.TRUNCATE,
 842        "UNION": TokenType.UNION,
 843        "UNKNOWN": TokenType.UNKNOWN,
 844        "UNNEST": TokenType.UNNEST,
 845        "UNPIVOT": TokenType.UNPIVOT,
 846        "UPDATE": TokenType.UPDATE,
 847        "USE": TokenType.USE,
 848        "USING": TokenType.USING,
 849        "UUID": TokenType.UUID,
 850        "VALUES": TokenType.VALUES,
 851        "VIEW": TokenType.VIEW,
 852        "VOLATILE": TokenType.VOLATILE,
 853        "WHEN": TokenType.WHEN,
 854        "WHERE": TokenType.WHERE,
 855        "WINDOW": TokenType.WINDOW,
 856        "WITH": TokenType.WITH,
 857        "APPLY": TokenType.APPLY,
 858        "ARRAY": TokenType.ARRAY,
 859        "BIT": TokenType.BIT,
 860        "BOOL": TokenType.BOOLEAN,
 861        "BOOLEAN": TokenType.BOOLEAN,
 862        "BYTE": TokenType.TINYINT,
 863        "MEDIUMINT": TokenType.MEDIUMINT,
 864        "INT1": TokenType.TINYINT,
 865        "TINYINT": TokenType.TINYINT,
 866        "INT16": TokenType.SMALLINT,
 867        "SHORT": TokenType.SMALLINT,
 868        "SMALLINT": TokenType.SMALLINT,
 869        "HUGEINT": TokenType.INT128,
 870        "UHUGEINT": TokenType.UINT128,
 871        "INT2": TokenType.SMALLINT,
 872        "INTEGER": TokenType.INT,
 873        "INT": TokenType.INT,
 874        "INT4": TokenType.INT,
 875        "INT32": TokenType.INT,
 876        "INT64": TokenType.BIGINT,
 877        "INT128": TokenType.INT128,
 878        "INT256": TokenType.INT256,
 879        "LONG": TokenType.BIGINT,
 880        "BIGINT": TokenType.BIGINT,
 881        "INT8": TokenType.TINYINT,
 882        "UINT": TokenType.UINT,
 883        "UINT128": TokenType.UINT128,
 884        "UINT256": TokenType.UINT256,
 885        "DEC": TokenType.DECIMAL,
 886        "DECIMAL": TokenType.DECIMAL,
 887        "DECIMAL32": TokenType.DECIMAL32,
 888        "DECIMAL64": TokenType.DECIMAL64,
 889        "DECIMAL128": TokenType.DECIMAL128,
 890        "DECIMAL256": TokenType.DECIMAL256,
 891        "BIGDECIMAL": TokenType.BIGDECIMAL,
 892        "BIGNUMERIC": TokenType.BIGDECIMAL,
 893        "LIST": TokenType.LIST,
 894        "MAP": TokenType.MAP,
 895        "NULLABLE": TokenType.NULLABLE,
 896        "NUMBER": TokenType.DECIMAL,
 897        "NUMERIC": TokenType.DECIMAL,
 898        "FIXED": TokenType.DECIMAL,
 899        "REAL": TokenType.FLOAT,
 900        "FLOAT": TokenType.FLOAT,
 901        "FLOAT4": TokenType.FLOAT,
 902        "FLOAT8": TokenType.DOUBLE,
 903        "DOUBLE": TokenType.DOUBLE,
 904        "DOUBLE PRECISION": TokenType.DOUBLE,
 905        "JSON": TokenType.JSON,
 906        "JSONB": TokenType.JSONB,
 907        "CHAR": TokenType.CHAR,
 908        "CHARACTER": TokenType.CHAR,
 909        "NCHAR": TokenType.NCHAR,
 910        "VARCHAR": TokenType.VARCHAR,
 911        "VARCHAR2": TokenType.VARCHAR,
 912        "NVARCHAR": TokenType.NVARCHAR,
 913        "NVARCHAR2": TokenType.NVARCHAR,
 914        "BPCHAR": TokenType.BPCHAR,
 915        "STR": TokenType.TEXT,
 916        "STRING": TokenType.TEXT,
 917        "TEXT": TokenType.TEXT,
 918        "LONGTEXT": TokenType.LONGTEXT,
 919        "MEDIUMTEXT": TokenType.MEDIUMTEXT,
 920        "TINYTEXT": TokenType.TINYTEXT,
 921        "CLOB": TokenType.TEXT,
 922        "LONGVARCHAR": TokenType.TEXT,
 923        "BINARY": TokenType.BINARY,
 924        "BLOB": TokenType.VARBINARY,
 925        "LONGBLOB": TokenType.LONGBLOB,
 926        "MEDIUMBLOB": TokenType.MEDIUMBLOB,
 927        "TINYBLOB": TokenType.TINYBLOB,
 928        "BYTEA": TokenType.VARBINARY,
 929        "VARBINARY": TokenType.VARBINARY,
 930        "TIME": TokenType.TIME,
 931        "TIMETZ": TokenType.TIMETZ,
 932        "TIMESTAMP": TokenType.TIMESTAMP,
 933        "TIMESTAMPTZ": TokenType.TIMESTAMPTZ,
 934        "TIMESTAMPLTZ": TokenType.TIMESTAMPLTZ,
 935        "TIMESTAMP_LTZ": TokenType.TIMESTAMPLTZ,
 936        "TIMESTAMPNTZ": TokenType.TIMESTAMPNTZ,
 937        "TIMESTAMP_NTZ": TokenType.TIMESTAMPNTZ,
 938        "DATE": TokenType.DATE,
 939        "DATETIME": TokenType.DATETIME,
 940        "INT4RANGE": TokenType.INT4RANGE,
 941        "INT4MULTIRANGE": TokenType.INT4MULTIRANGE,
 942        "INT8RANGE": TokenType.INT8RANGE,
 943        "INT8MULTIRANGE": TokenType.INT8MULTIRANGE,
 944        "NUMRANGE": TokenType.NUMRANGE,
 945        "NUMMULTIRANGE": TokenType.NUMMULTIRANGE,
 946        "TSRANGE": TokenType.TSRANGE,
 947        "TSMULTIRANGE": TokenType.TSMULTIRANGE,
 948        "TSTZRANGE": TokenType.TSTZRANGE,
 949        "TSTZMULTIRANGE": TokenType.TSTZMULTIRANGE,
 950        "DATERANGE": TokenType.DATERANGE,
 951        "DATEMULTIRANGE": TokenType.DATEMULTIRANGE,
 952        "UNIQUE": TokenType.UNIQUE,
 953        "VECTOR": TokenType.VECTOR,
 954        "STRUCT": TokenType.STRUCT,
 955        "SEQUENCE": TokenType.SEQUENCE,
 956        "VARIANT": TokenType.VARIANT,
 957        "ALTER": TokenType.ALTER,
 958        "ANALYZE": TokenType.ANALYZE,
 959        "CALL": TokenType.COMMAND,
 960        "COMMENT": TokenType.COMMENT,
 961        "EXPLAIN": TokenType.COMMAND,
 962        "GRANT": TokenType.GRANT,
 963        "OPTIMIZE": TokenType.COMMAND,
 964        "PREPARE": TokenType.COMMAND,
 965        "VACUUM": TokenType.COMMAND,
 966        "USER-DEFINED": TokenType.USERDEFINED,
 967        "FOR VERSION": TokenType.VERSION_SNAPSHOT,
 968        "FOR TIMESTAMP": TokenType.TIMESTAMP_SNAPSHOT,
 969    }
 970
 971    WHITE_SPACE: t.Dict[t.Optional[str], TokenType] = {
 972        " ": TokenType.SPACE,
 973        "\t": TokenType.SPACE,
 974        "\n": TokenType.BREAK,
 975        "\r": TokenType.BREAK,
 976    }
 977
 978    COMMANDS = {
 979        TokenType.COMMAND,
 980        TokenType.EXECUTE,
 981        TokenType.FETCH,
 982        TokenType.SHOW,
 983        TokenType.RENAME,
 984    }
 985
 986    COMMAND_PREFIX_TOKENS = {TokenType.SEMICOLON, TokenType.BEGIN}
 987
 988    # Handle numeric literals like in hive (3L = BIGINT)
 989    NUMERIC_LITERALS: t.Dict[str, str] = {}
 990
 991    COMMENTS = ["--", ("/*", "*/")]
 992
 993    __slots__ = (
 994        "sql",
 995        "size",
 996        "tokens",
 997        "dialect",
 998        "use_rs_tokenizer",
 999        "_start",
1000        "_current",
1001        "_line",
1002        "_col",
1003        "_comments",
1004        "_char",
1005        "_end",
1006        "_peek",
1007        "_prev_token_line",
1008        "_rs_dialect_settings",
1009    )
1010
1011    def __init__(
1012        self, dialect: DialectType = None, use_rs_tokenizer: t.Optional[bool] = None
1013    ) -> None:
1014        from sqlglot.dialects import Dialect
1015
1016        self.dialect = Dialect.get_or_raise(dialect)
1017
1018        # initialize `use_rs_tokenizer`, and allow it to be overwritten per Tokenizer instance
1019        self.use_rs_tokenizer = (
1020            use_rs_tokenizer if use_rs_tokenizer is not None else USE_RS_TOKENIZER
1021        )
1022
1023        if self.use_rs_tokenizer:
1024            self._rs_dialect_settings = RsTokenizerDialectSettings(
1025                unescaped_sequences=self.dialect.UNESCAPED_SEQUENCES,
1026                identifiers_can_start_with_digit=self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT,
1027                numbers_can_be_underscore_separated=self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED,
1028            )
1029
1030        self.reset()
1031
1032    def reset(self) -> None:
1033        self.sql = ""
1034        self.size = 0
1035        self.tokens: t.List[Token] = []
1036        self._start = 0
1037        self._current = 0
1038        self._line = 1
1039        self._col = 0
1040        self._comments: t.List[str] = []
1041
1042        self._char = ""
1043        self._end = False
1044        self._peek = ""
1045        self._prev_token_line = -1
1046
1047    def tokenize(self, sql: str) -> t.List[Token]:
1048        """Returns a list of tokens corresponding to the SQL string `sql`."""
1049        if self.use_rs_tokenizer:
1050            return self.tokenize_rs(sql)
1051
1052        self.reset()
1053        self.sql = sql
1054        self.size = len(sql)
1055
1056        try:
1057            self._scan()
1058        except Exception as e:
1059            start = max(self._current - 50, 0)
1060            end = min(self._current + 50, self.size - 1)
1061            context = self.sql[start:end]
1062            raise TokenError(f"Error tokenizing '{context}'") from e
1063
1064        return self.tokens
1065
1066    def _scan(self, until: t.Optional[t.Callable] = None) -> None:
1067        while self.size and not self._end:
1068            current = self._current
1069
1070            # Skip spaces here rather than iteratively calling advance() for performance reasons
1071            while current < self.size:
1072                char = self.sql[current]
1073
1074                if char.isspace() and (char == " " or char == "\t"):
1075                    current += 1
1076                else:
1077                    break
1078
1079            offset = current - self._current if current > self._current else 1
1080
1081            self._start = current
1082            self._advance(offset)
1083
1084            if not self._char.isspace():
1085                if self._char.isdigit():
1086                    self._scan_number()
1087                elif self._char in self._IDENTIFIERS:
1088                    self._scan_identifier(self._IDENTIFIERS[self._char])
1089                else:
1090                    self._scan_keywords()
1091
1092            if until and until():
1093                break
1094
1095        if self.tokens and self._comments:
1096            self.tokens[-1].comments.extend(self._comments)
1097
1098    def _chars(self, size: int) -> str:
1099        if size == 1:
1100            return self._char
1101
1102        start = self._current - 1
1103        end = start + size
1104
1105        return self.sql[start:end] if end <= self.size else ""
1106
1107    def _advance(self, i: int = 1, alnum: bool = False) -> None:
1108        if self.WHITE_SPACE.get(self._char) is TokenType.BREAK:
1109            # Ensures we don't count an extra line if we get a \r\n line break sequence
1110            if not (self._char == "\r" and self._peek == "\n"):
1111                self._col = 1
1112                self._line += 1
1113        else:
1114            self._col += i
1115
1116        self._current += i
1117        self._end = self._current >= self.size
1118        self._char = self.sql[self._current - 1]
1119        self._peek = "" if self._end else self.sql[self._current]
1120
1121        if alnum and self._char.isalnum():
1122            # Here we use local variables instead of attributes for better performance
1123            _col = self._col
1124            _current = self._current
1125            _end = self._end
1126            _peek = self._peek
1127
1128            while _peek.isalnum():
1129                _col += 1
1130                _current += 1
1131                _end = _current >= self.size
1132                _peek = "" if _end else self.sql[_current]
1133
1134            self._col = _col
1135            self._current = _current
1136            self._end = _end
1137            self._peek = _peek
1138            self._char = self.sql[_current - 1]
1139
1140    @property
1141    def _text(self) -> str:
1142        return self.sql[self._start : self._current]
1143
1144    def _add(self, token_type: TokenType, text: t.Optional[str] = None) -> None:
1145        self._prev_token_line = self._line
1146
1147        if self._comments and token_type == TokenType.SEMICOLON and self.tokens:
1148            self.tokens[-1].comments.extend(self._comments)
1149            self._comments = []
1150
1151        self.tokens.append(
1152            Token(
1153                token_type,
1154                text=self._text if text is None else text,
1155                line=self._line,
1156                col=self._col,
1157                start=self._start,
1158                end=self._current - 1,
1159                comments=self._comments,
1160            )
1161        )
1162        self._comments = []
1163
1164        # If we have either a semicolon or a begin token before the command's token, we'll parse
1165        # whatever follows the command's token as a string
1166        if (
1167            token_type in self.COMMANDS
1168            and self._peek != ";"
1169            and (len(self.tokens) == 1 or self.tokens[-2].token_type in self.COMMAND_PREFIX_TOKENS)
1170        ):
1171            start = self._current
1172            tokens = len(self.tokens)
1173            self._scan(lambda: self._peek == ";")
1174            self.tokens = self.tokens[:tokens]
1175            text = self.sql[start : self._current].strip()
1176            if text:
1177                self._add(TokenType.STRING, text)
1178
1179    def _scan_keywords(self) -> None:
1180        size = 0
1181        word = None
1182        chars = self._text
1183        char = chars
1184        prev_space = False
1185        skip = False
1186        trie = self._KEYWORD_TRIE
1187        single_token = char in self.SINGLE_TOKENS
1188
1189        while chars:
1190            if skip:
1191                result = TrieResult.PREFIX
1192            else:
1193                result, trie = in_trie(trie, char.upper())
1194
1195            if result == TrieResult.FAILED:
1196                break
1197            if result == TrieResult.EXISTS:
1198                word = chars
1199
1200            end = self._current + size
1201            size += 1
1202
1203            if end < self.size:
1204                char = self.sql[end]
1205                single_token = single_token or char in self.SINGLE_TOKENS
1206                is_space = char.isspace()
1207
1208                if not is_space or not prev_space:
1209                    if is_space:
1210                        char = " "
1211                    chars += char
1212                    prev_space = is_space
1213                    skip = False
1214                else:
1215                    skip = True
1216            else:
1217                char = ""
1218                break
1219
1220        if word:
1221            if self._scan_string(word):
1222                return
1223            if self._scan_comment(word):
1224                return
1225            if prev_space or single_token or not char:
1226                self._advance(size - 1)
1227                word = word.upper()
1228                self._add(self.KEYWORDS[word], text=word)
1229                return
1230
1231        if self._char in self.SINGLE_TOKENS:
1232            self._add(self.SINGLE_TOKENS[self._char], text=self._char)
1233            return
1234
1235        self._scan_var()
1236
1237    def _scan_comment(self, comment_start: str) -> bool:
1238        if comment_start not in self._COMMENTS:
1239            return False
1240
1241        comment_start_line = self._line
1242        comment_start_size = len(comment_start)
1243        comment_end = self._COMMENTS[comment_start]
1244
1245        if comment_end:
1246            # Skip the comment's start delimiter
1247            self._advance(comment_start_size)
1248
1249            comment_count = 1
1250            comment_end_size = len(comment_end)
1251
1252            while not self._end:
1253                if self._chars(comment_end_size) == comment_end:
1254                    comment_count -= 1
1255                    if not comment_count:
1256                        break
1257
1258                self._advance(alnum=True)
1259
1260                # Nested comments are allowed by some dialects, e.g. databricks, duckdb, postgres
1261                if (
1262                    self.NESTED_COMMENTS
1263                    and not self._end
1264                    and self._chars(comment_end_size) == comment_start
1265                ):
1266                    self._advance(comment_start_size)
1267                    comment_count += 1
1268
1269            self._comments.append(self._text[comment_start_size : -comment_end_size + 1])
1270            self._advance(comment_end_size - 1)
1271        else:
1272            while not self._end and self.WHITE_SPACE.get(self._peek) is not TokenType.BREAK:
1273                self._advance(alnum=True)
1274            self._comments.append(self._text[comment_start_size:])
1275
1276        if (
1277            comment_start == self.HINT_START
1278            and self.tokens
1279            and self.tokens[-1].token_type in self.TOKENS_PRECEDING_HINT
1280        ):
1281            self._add(TokenType.HINT)
1282
1283        # Leading comment is attached to the succeeding token, whilst trailing comment to the preceding.
1284        # Multiple consecutive comments are preserved by appending them to the current comments list.
1285        if comment_start_line == self._prev_token_line:
1286            self.tokens[-1].comments.extend(self._comments)
1287            self._comments = []
1288            self._prev_token_line = self._line
1289
1290        return True
1291
1292    def _scan_number(self) -> None:
1293        if self._char == "0":
1294            peek = self._peek.upper()
1295            if peek == "B":
1296                return self._scan_bits() if self.BIT_STRINGS else self._add(TokenType.NUMBER)
1297            elif peek == "X":
1298                return self._scan_hex() if self.HEX_STRINGS else self._add(TokenType.NUMBER)
1299
1300        decimal = False
1301        scientific = 0
1302
1303        while True:
1304            if self._peek.isdigit():
1305                self._advance()
1306            elif self._peek == "." and not decimal:
1307                if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER:
1308                    return self._add(TokenType.NUMBER)
1309                decimal = True
1310                self._advance()
1311            elif self._peek in ("-", "+") and scientific == 1:
1312                scientific += 1
1313                self._advance()
1314            elif self._peek.upper() == "E" and not scientific:
1315                scientific += 1
1316                self._advance()
1317            elif self._peek.isidentifier():
1318                number_text = self._text
1319                literal = ""
1320
1321                while self._peek.strip() and self._peek not in self.SINGLE_TOKENS:
1322                    literal += self._peek
1323                    self._advance()
1324
1325                token_type = self.KEYWORDS.get(self.NUMERIC_LITERALS.get(literal.upper(), ""))
1326
1327                if token_type:
1328                    self._add(TokenType.NUMBER, number_text)
1329                    self._add(TokenType.DCOLON, "::")
1330                    return self._add(token_type, literal)
1331                else:
1332                    replaced = literal.replace("_", "")
1333                    if self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED and replaced.isdigit():
1334                        return self._add(TokenType.NUMBER, number_text + replaced)
1335                    if self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT:
1336                        return self._add(TokenType.VAR)
1337
1338                self._advance(-len(literal))
1339                return self._add(TokenType.NUMBER, number_text)
1340            else:
1341                return self._add(TokenType.NUMBER)
1342
1343    def _scan_bits(self) -> None:
1344        self._advance()
1345        value = self._extract_value()
1346        try:
1347            # If `value` can't be converted to a binary, fallback to tokenizing it as an identifier
1348            int(value, 2)
1349            self._add(TokenType.BIT_STRING, value[2:])  # Drop the 0b
1350        except ValueError:
1351            self._add(TokenType.IDENTIFIER)
1352
1353    def _scan_hex(self) -> None:
1354        self._advance()
1355        value = self._extract_value()
1356        try:
1357            # If `value` can't be converted to a hex, fallback to tokenizing it as an identifier
1358            int(value, 16)
1359            self._add(TokenType.HEX_STRING, value[2:])  # Drop the 0x
1360        except ValueError:
1361            self._add(TokenType.IDENTIFIER)
1362
1363    def _extract_value(self) -> str:
1364        while True:
1365            char = self._peek.strip()
1366            if char and char not in self.SINGLE_TOKENS:
1367                self._advance(alnum=True)
1368            else:
1369                break
1370
1371        return self._text
1372
1373    def _scan_string(self, start: str) -> bool:
1374        base = None
1375        token_type = TokenType.STRING
1376
1377        if start in self._QUOTES:
1378            end = self._QUOTES[start]
1379        elif start in self._FORMAT_STRINGS:
1380            end, token_type = self._FORMAT_STRINGS[start]
1381
1382            if token_type == TokenType.HEX_STRING:
1383                base = 16
1384            elif token_type == TokenType.BIT_STRING:
1385                base = 2
1386            elif token_type == TokenType.HEREDOC_STRING:
1387                self._advance()
1388
1389                if self._char == end:
1390                    tag = ""
1391                else:
1392                    tag = self._extract_string(
1393                        end,
1394                        raw_string=True,
1395                        raise_unmatched=not self.HEREDOC_TAG_IS_IDENTIFIER,
1396                    )
1397
1398                if tag and self.HEREDOC_TAG_IS_IDENTIFIER and (self._end or not tag.isidentifier()):
1399                    if not self._end:
1400                        self._advance(-1)
1401
1402                    self._advance(-len(tag))
1403                    self._add(self.HEREDOC_STRING_ALTERNATIVE)
1404                    return True
1405
1406                end = f"{start}{tag}{end}"
1407        else:
1408            return False
1409
1410        self._advance(len(start))
1411        text = self._extract_string(end, raw_string=token_type == TokenType.RAW_STRING)
1412
1413        if base:
1414            try:
1415                int(text, base)
1416            except Exception:
1417                raise TokenError(
1418                    f"Numeric string contains invalid characters from {self._line}:{self._start}"
1419                )
1420
1421        self._add(token_type, text)
1422        return True
1423
1424    def _scan_identifier(self, identifier_end: str) -> None:
1425        self._advance()
1426        text = self._extract_string(
1427            identifier_end, escapes=self._IDENTIFIER_ESCAPES | {identifier_end}
1428        )
1429        self._add(TokenType.IDENTIFIER, text)
1430
1431    def _scan_var(self) -> None:
1432        while True:
1433            char = self._peek.strip()
1434            if char and (char in self.VAR_SINGLE_TOKENS or char not in self.SINGLE_TOKENS):
1435                self._advance(alnum=True)
1436            else:
1437                break
1438
1439        self._add(
1440            TokenType.VAR
1441            if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER
1442            else self.KEYWORDS.get(self._text.upper(), TokenType.VAR)
1443        )
1444
1445    def _extract_string(
1446        self,
1447        delimiter: str,
1448        escapes: t.Optional[t.Set[str]] = None,
1449        raw_string: bool = False,
1450        raise_unmatched: bool = True,
1451    ) -> str:
1452        text = ""
1453        delim_size = len(delimiter)
1454        escapes = self._STRING_ESCAPES if escapes is None else escapes
1455
1456        while True:
1457            if (
1458                not raw_string
1459                and self.dialect.UNESCAPED_SEQUENCES
1460                and self._peek
1461                and self._char in self.STRING_ESCAPES
1462            ):
1463                unescaped_sequence = self.dialect.UNESCAPED_SEQUENCES.get(self._char + self._peek)
1464                if unescaped_sequence:
1465                    self._advance(2)
1466                    text += unescaped_sequence
1467                    continue
1468            if (
1469                (self.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS or not raw_string)
1470                and self._char in escapes
1471                and (self._peek == delimiter or self._peek in escapes)
1472                and (self._char not in self._QUOTES or self._char == self._peek)
1473            ):
1474                if self._peek == delimiter:
1475                    text += self._peek
1476                else:
1477                    text += self._char + self._peek
1478
1479                if self._current + 1 < self.size:
1480                    self._advance(2)
1481                else:
1482                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._current}")
1483            else:
1484                if self._chars(delim_size) == delimiter:
1485                    if delim_size > 1:
1486                        self._advance(delim_size - 1)
1487                    break
1488
1489                if self._end:
1490                    if not raise_unmatched:
1491                        return text + self._char
1492
1493                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._start}")
1494
1495                current = self._current - 1
1496                self._advance(alnum=True)
1497                text += self.sql[current : self._current - 1]
1498
1499        return text
1500
1501    def tokenize_rs(self, sql: str) -> t.List[Token]:
1502        if not self._RS_TOKENIZER:
1503            raise SqlglotError("Rust tokenizer is not available")
1504
1505        try:
1506            tokens = self._RS_TOKENIZER.tokenize(sql, self._rs_dialect_settings)
1507            for token in tokens:
1508                token.token_type = _ALL_TOKEN_TYPES[token.token_type_index]
1509            return tokens
1510        except Exception as e:
1511            raise TokenError(str(e))