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    DCOLONDOLLAR = auto()
  43    DCOLONPERCENT = auto()
  44    DQMARK = auto()
  45    SEMICOLON = auto()
  46    STAR = auto()
  47    BACKSLASH = auto()
  48    SLASH = auto()
  49    LT = auto()
  50    LTE = auto()
  51    GT = auto()
  52    GTE = auto()
  53    NOT = auto()
  54    EQ = auto()
  55    NEQ = auto()
  56    NULLSAFE_EQ = auto()
  57    COLON_EQ = auto()
  58    COLON_GT = auto()
  59    NCOLON_GT = auto()
  60    AND = auto()
  61    OR = auto()
  62    AMP = auto()
  63    DPIPE = auto()
  64    PIPE_GT = auto()
  65    PIPE = auto()
  66    PIPE_SLASH = auto()
  67    DPIPE_SLASH = auto()
  68    CARET = auto()
  69    CARET_AT = auto()
  70    TILDA = auto()
  71    ARROW = auto()
  72    DARROW = auto()
  73    FARROW = auto()
  74    HASH = auto()
  75    HASH_ARROW = auto()
  76    DHASH_ARROW = auto()
  77    LR_ARROW = auto()
  78    DAT = auto()
  79    LT_AT = auto()
  80    AT_GT = auto()
  81    DOLLAR = auto()
  82    PARAMETER = auto()
  83    SESSION = auto()
  84    SESSION_PARAMETER = auto()
  85    DAMP = auto()
  86    XOR = auto()
  87    DSTAR = auto()
  88    QMARK_AMP = auto()
  89    QMARK_PIPE = auto()
  90    HASH_DASH = auto()
  91
  92    URI_START = auto()
  93
  94    BLOCK_START = auto()
  95    BLOCK_END = auto()
  96
  97    SPACE = auto()
  98    BREAK = auto()
  99
 100    STRING = auto()
 101    NUMBER = auto()
 102    IDENTIFIER = auto()
 103    DATABASE = auto()
 104    COLUMN = auto()
 105    COLUMN_DEF = auto()
 106    SCHEMA = auto()
 107    TABLE = auto()
 108    WAREHOUSE = auto()
 109    STAGE = auto()
 110    STREAMLIT = auto()
 111    VAR = auto()
 112    BIT_STRING = auto()
 113    HEX_STRING = auto()
 114    BYTE_STRING = auto()
 115    NATIONAL_STRING = auto()
 116    RAW_STRING = auto()
 117    HEREDOC_STRING = auto()
 118    UNICODE_STRING = auto()
 119
 120    # types
 121    BIT = auto()
 122    BOOLEAN = auto()
 123    TINYINT = auto()
 124    UTINYINT = auto()
 125    SMALLINT = auto()
 126    USMALLINT = auto()
 127    MEDIUMINT = auto()
 128    UMEDIUMINT = auto()
 129    INT = auto()
 130    UINT = auto()
 131    BIGINT = auto()
 132    UBIGINT = auto()
 133    INT128 = auto()
 134    UINT128 = auto()
 135    INT256 = auto()
 136    UINT256 = auto()
 137    FLOAT = auto()
 138    DOUBLE = auto()
 139    UDOUBLE = auto()
 140    DECIMAL = auto()
 141    DECIMAL32 = auto()
 142    DECIMAL64 = auto()
 143    DECIMAL128 = auto()
 144    DECIMAL256 = auto()
 145    UDECIMAL = auto()
 146    BIGDECIMAL = auto()
 147    CHAR = auto()
 148    NCHAR = auto()
 149    VARCHAR = auto()
 150    NVARCHAR = auto()
 151    BPCHAR = auto()
 152    TEXT = auto()
 153    MEDIUMTEXT = auto()
 154    LONGTEXT = auto()
 155    BLOB = auto()
 156    MEDIUMBLOB = auto()
 157    LONGBLOB = auto()
 158    TINYBLOB = auto()
 159    TINYTEXT = auto()
 160    NAME = auto()
 161    BINARY = auto()
 162    VARBINARY = auto()
 163    JSON = auto()
 164    JSONB = auto()
 165    TIME = auto()
 166    TIMETZ = auto()
 167    TIMESTAMP = auto()
 168    TIMESTAMPTZ = auto()
 169    TIMESTAMPLTZ = auto()
 170    TIMESTAMPNTZ = auto()
 171    TIMESTAMP_S = auto()
 172    TIMESTAMP_MS = auto()
 173    TIMESTAMP_NS = auto()
 174    DATETIME = auto()
 175    DATETIME2 = auto()
 176    DATETIME64 = auto()
 177    SMALLDATETIME = auto()
 178    DATE = auto()
 179    DATE32 = auto()
 180    INT4RANGE = auto()
 181    INT4MULTIRANGE = auto()
 182    INT8RANGE = auto()
 183    INT8MULTIRANGE = auto()
 184    NUMRANGE = auto()
 185    NUMMULTIRANGE = auto()
 186    TSRANGE = auto()
 187    TSMULTIRANGE = auto()
 188    TSTZRANGE = auto()
 189    TSTZMULTIRANGE = auto()
 190    DATERANGE = auto()
 191    DATEMULTIRANGE = auto()
 192    UUID = auto()
 193    GEOGRAPHY = auto()
 194    GEOGRAPHYPOINT = auto()
 195    NULLABLE = auto()
 196    GEOMETRY = auto()
 197    POINT = auto()
 198    RING = auto()
 199    LINESTRING = auto()
 200    MULTILINESTRING = auto()
 201    POLYGON = auto()
 202    MULTIPOLYGON = auto()
 203    HLLSKETCH = auto()
 204    HSTORE = auto()
 205    SUPER = auto()
 206    SERIAL = auto()
 207    SMALLSERIAL = auto()
 208    BIGSERIAL = auto()
 209    XML = auto()
 210    YEAR = auto()
 211    USERDEFINED = auto()
 212    MONEY = auto()
 213    SMALLMONEY = auto()
 214    ROWVERSION = auto()
 215    IMAGE = auto()
 216    VARIANT = auto()
 217    OBJECT = auto()
 218    INET = auto()
 219    IPADDRESS = auto()
 220    IPPREFIX = auto()
 221    IPV4 = auto()
 222    IPV6 = auto()
 223    ENUM = auto()
 224    ENUM8 = auto()
 225    ENUM16 = auto()
 226    FIXEDSTRING = auto()
 227    LOWCARDINALITY = auto()
 228    NESTED = auto()
 229    AGGREGATEFUNCTION = auto()
 230    SIMPLEAGGREGATEFUNCTION = auto()
 231    TDIGEST = auto()
 232    UNKNOWN = auto()
 233    VECTOR = auto()
 234    DYNAMIC = auto()
 235    VOID = auto()
 236
 237    # keywords
 238    ALIAS = auto()
 239    ALTER = auto()
 240    ALL = auto()
 241    ANTI = auto()
 242    ANY = auto()
 243    APPLY = auto()
 244    ARRAY = auto()
 245    ASC = auto()
 246    ASOF = auto()
 247    ATTACH = auto()
 248    AUTO_INCREMENT = auto()
 249    BEGIN = auto()
 250    BETWEEN = auto()
 251    BULK_COLLECT_INTO = auto()
 252    CACHE = auto()
 253    CASE = auto()
 254    CHARACTER_SET = auto()
 255    CLUSTER_BY = auto()
 256    COLLATE = auto()
 257    COMMAND = auto()
 258    COMMENT = auto()
 259    COMMIT = auto()
 260    CONNECT_BY = auto()
 261    CONSTRAINT = auto()
 262    COPY = auto()
 263    CREATE = auto()
 264    CROSS = auto()
 265    CUBE = auto()
 266    CURRENT_DATE = auto()
 267    CURRENT_DATETIME = auto()
 268    CURRENT_SCHEMA = auto()
 269    CURRENT_TIME = auto()
 270    CURRENT_TIMESTAMP = auto()
 271    CURRENT_USER = auto()
 272    DECLARE = auto()
 273    DEFAULT = auto()
 274    DELETE = auto()
 275    DESC = auto()
 276    DESCRIBE = auto()
 277    DETACH = auto()
 278    DICTIONARY = auto()
 279    DISTINCT = auto()
 280    DISTRIBUTE_BY = auto()
 281    DIV = auto()
 282    DROP = auto()
 283    ELSE = auto()
 284    END = auto()
 285    ESCAPE = auto()
 286    EXCEPT = auto()
 287    EXECUTE = auto()
 288    EXISTS = auto()
 289    FALSE = auto()
 290    FETCH = auto()
 291    FILE_FORMAT = auto()
 292    FILTER = auto()
 293    FINAL = auto()
 294    FIRST = auto()
 295    FOR = auto()
 296    FORCE = auto()
 297    FOREIGN_KEY = auto()
 298    FORMAT = auto()
 299    FROM = auto()
 300    FULL = auto()
 301    FUNCTION = auto()
 302    GET = auto()
 303    GLOB = auto()
 304    GLOBAL = auto()
 305    GRANT = auto()
 306    GROUP_BY = auto()
 307    GROUPING_SETS = auto()
 308    HAVING = auto()
 309    HINT = auto()
 310    IGNORE = auto()
 311    ILIKE = auto()
 312    IN = auto()
 313    INDEX = auto()
 314    INNER = auto()
 315    INSERT = auto()
 316    INTERSECT = auto()
 317    INTERVAL = auto()
 318    INTO = auto()
 319    INTRODUCER = auto()
 320    IRLIKE = auto()
 321    IS = auto()
 322    ISNULL = auto()
 323    JOIN = auto()
 324    JOIN_MARKER = auto()
 325    KEEP = auto()
 326    KEY = auto()
 327    KILL = auto()
 328    LANGUAGE = auto()
 329    LATERAL = auto()
 330    LEFT = auto()
 331    LIKE = auto()
 332    LIMIT = auto()
 333    LIST = auto()
 334    LOAD = auto()
 335    LOCK = auto()
 336    MAP = auto()
 337    MATCH_CONDITION = auto()
 338    MATCH_RECOGNIZE = auto()
 339    MEMBER_OF = auto()
 340    MERGE = auto()
 341    MOD = auto()
 342    MODEL = auto()
 343    NATURAL = auto()
 344    NEXT = auto()
 345    NOTHING = auto()
 346    NOTNULL = auto()
 347    NULL = auto()
 348    OBJECT_IDENTIFIER = auto()
 349    OFFSET = auto()
 350    ON = auto()
 351    ONLY = auto()
 352    OPERATOR = auto()
 353    ORDER_BY = auto()
 354    ORDER_SIBLINGS_BY = auto()
 355    ORDERED = auto()
 356    ORDINALITY = auto()
 357    OUTER = auto()
 358    OVER = auto()
 359    OVERLAPS = auto()
 360    OVERWRITE = auto()
 361    PARTITION = auto()
 362    PARTITION_BY = auto()
 363    PERCENT = auto()
 364    PIVOT = auto()
 365    PLACEHOLDER = auto()
 366    POSITIONAL = auto()
 367    PRAGMA = auto()
 368    PREWHERE = auto()
 369    PRIMARY_KEY = auto()
 370    PROCEDURE = auto()
 371    PROPERTIES = auto()
 372    PSEUDO_TYPE = auto()
 373    PUT = auto()
 374    QUALIFY = auto()
 375    QUOTE = auto()
 376    RANGE = auto()
 377    RECURSIVE = auto()
 378    REFRESH = auto()
 379    RENAME = auto()
 380    REPLACE = auto()
 381    RETURNING = auto()
 382    REVOKE = auto()
 383    REFERENCES = auto()
 384    RIGHT = auto()
 385    RLIKE = auto()
 386    ROLLBACK = auto()
 387    ROLLUP = auto()
 388    ROW = auto()
 389    ROWS = auto()
 390    SELECT = auto()
 391    SEMI = auto()
 392    SEPARATOR = auto()
 393    SEQUENCE = auto()
 394    SERDE_PROPERTIES = auto()
 395    SET = auto()
 396    SETTINGS = auto()
 397    SHOW = auto()
 398    SIMILAR_TO = auto()
 399    SOME = auto()
 400    SORT_BY = auto()
 401    START_WITH = auto()
 402    STORAGE_INTEGRATION = auto()
 403    STRAIGHT_JOIN = auto()
 404    STRUCT = auto()
 405    SUMMARIZE = auto()
 406    TABLE_SAMPLE = auto()
 407    TAG = auto()
 408    TEMPORARY = auto()
 409    TOP = auto()
 410    THEN = auto()
 411    TRUE = auto()
 412    TRUNCATE = auto()
 413    UNCACHE = auto()
 414    UNION = auto()
 415    UNNEST = auto()
 416    UNPIVOT = auto()
 417    UPDATE = auto()
 418    USE = auto()
 419    USING = auto()
 420    VALUES = auto()
 421    VIEW = auto()
 422    SEMANTIC_VIEW = auto()
 423    VOLATILE = auto()
 424    WHEN = auto()
 425    WHERE = auto()
 426    WINDOW = auto()
 427    WITH = auto()
 428    UNIQUE = auto()
 429    UTC_DATE = auto()
 430    UTC_TIME = auto()
 431    UTC_TIMESTAMP = auto()
 432    VERSION_SNAPSHOT = auto()
 433    TIMESTAMP_SNAPSHOT = auto()
 434    OPTION = auto()
 435    SINK = auto()
 436    SOURCE = auto()
 437    ANALYZE = auto()
 438    NAMESPACE = auto()
 439    EXPORT = auto()
 440
 441    # sentinel
 442    HIVE_TOKEN_STREAM = auto()
 443
 444
 445_ALL_TOKEN_TYPES = list(TokenType)
 446_TOKEN_TYPE_TO_INDEX = {token_type: i for i, token_type in enumerate(_ALL_TOKEN_TYPES)}
 447
 448
 449class Token:
 450    __slots__ = ("token_type", "text", "line", "col", "start", "end", "comments")
 451
 452    @classmethod
 453    def number(cls, number: int) -> Token:
 454        """Returns a NUMBER token with `number` as its text."""
 455        return cls(TokenType.NUMBER, str(number))
 456
 457    @classmethod
 458    def string(cls, string: str) -> Token:
 459        """Returns a STRING token with `string` as its text."""
 460        return cls(TokenType.STRING, string)
 461
 462    @classmethod
 463    def identifier(cls, identifier: str) -> Token:
 464        """Returns an IDENTIFIER token with `identifier` as its text."""
 465        return cls(TokenType.IDENTIFIER, identifier)
 466
 467    @classmethod
 468    def var(cls, var: str) -> Token:
 469        """Returns an VAR token with `var` as its text."""
 470        return cls(TokenType.VAR, var)
 471
 472    def __init__(
 473        self,
 474        token_type: TokenType,
 475        text: str,
 476        line: int = 1,
 477        col: int = 1,
 478        start: int = 0,
 479        end: int = 0,
 480        comments: t.Optional[t.List[str]] = None,
 481    ) -> None:
 482        """Token initializer.
 483
 484        Args:
 485            token_type: The TokenType Enum.
 486            text: The text of the token.
 487            line: The line that the token ends on.
 488            col: The column that the token ends on.
 489            start: The start index of the token.
 490            end: The ending index of the token.
 491            comments: The comments to attach to the token.
 492        """
 493        self.token_type = token_type
 494        self.text = text
 495        self.line = line
 496        self.col = col
 497        self.start = start
 498        self.end = end
 499        self.comments = [] if comments is None else comments
 500
 501    def __repr__(self) -> str:
 502        attributes = ", ".join(f"{k}: {getattr(self, k)}" for k in self.__slots__)
 503        return f"<Token {attributes}>"
 504
 505
 506class _Tokenizer(type):
 507    def __new__(cls, clsname, bases, attrs):
 508        klass = super().__new__(cls, clsname, bases, attrs)
 509
 510        def _convert_quotes(arr: t.List[str | t.Tuple[str, str]]) -> t.Dict[str, str]:
 511            return dict(
 512                (item, item) if isinstance(item, str) else (item[0], item[1]) for item in arr
 513            )
 514
 515        def _quotes_to_format(
 516            token_type: TokenType, arr: t.List[str | t.Tuple[str, str]]
 517        ) -> t.Dict[str, t.Tuple[str, TokenType]]:
 518            return {k: (v, token_type) for k, v in _convert_quotes(arr).items()}
 519
 520        klass._QUOTES = _convert_quotes(klass.QUOTES)
 521        klass._IDENTIFIERS = _convert_quotes(klass.IDENTIFIERS)
 522
 523        klass._FORMAT_STRINGS = {
 524            **{
 525                p + s: (e, TokenType.NATIONAL_STRING)
 526                for s, e in klass._QUOTES.items()
 527                for p in ("n", "N")
 528            },
 529            **_quotes_to_format(TokenType.BIT_STRING, klass.BIT_STRINGS),
 530            **_quotes_to_format(TokenType.BYTE_STRING, klass.BYTE_STRINGS),
 531            **_quotes_to_format(TokenType.HEX_STRING, klass.HEX_STRINGS),
 532            **_quotes_to_format(TokenType.RAW_STRING, klass.RAW_STRINGS),
 533            **_quotes_to_format(TokenType.HEREDOC_STRING, klass.HEREDOC_STRINGS),
 534            **_quotes_to_format(TokenType.UNICODE_STRING, klass.UNICODE_STRINGS),
 535        }
 536
 537        klass._STRING_ESCAPES = set(klass.STRING_ESCAPES)
 538        klass._IDENTIFIER_ESCAPES = set(klass.IDENTIFIER_ESCAPES)
 539        klass._COMMENTS = {
 540            **dict(
 541                (comment, None) if isinstance(comment, str) else (comment[0], comment[1])
 542                for comment in klass.COMMENTS
 543            ),
 544            "{#": "#}",  # Ensure Jinja comments are tokenized correctly in all dialects
 545        }
 546        if klass.HINT_START in klass.KEYWORDS:
 547            klass._COMMENTS[klass.HINT_START] = "*/"
 548
 549        klass._KEYWORD_TRIE = new_trie(
 550            key.upper()
 551            for key in (
 552                *klass.KEYWORDS,
 553                *klass._COMMENTS,
 554                *klass._QUOTES,
 555                *klass._FORMAT_STRINGS,
 556            )
 557            if " " in key or any(single in key for single in klass.SINGLE_TOKENS)
 558        )
 559
 560        if USE_RS_TOKENIZER:
 561            settings = RsTokenizerSettings(
 562                white_space={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.WHITE_SPACE.items()},
 563                single_tokens={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.SINGLE_TOKENS.items()},
 564                keywords={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.KEYWORDS.items()},
 565                numeric_literals=klass.NUMERIC_LITERALS,
 566                identifiers=klass._IDENTIFIERS,
 567                identifier_escapes=klass._IDENTIFIER_ESCAPES,
 568                string_escapes=klass._STRING_ESCAPES,
 569                quotes=klass._QUOTES,
 570                format_strings={
 571                    k: (v1, _TOKEN_TYPE_TO_INDEX[v2])
 572                    for k, (v1, v2) in klass._FORMAT_STRINGS.items()
 573                },
 574                has_bit_strings=bool(klass.BIT_STRINGS),
 575                has_hex_strings=bool(klass.HEX_STRINGS),
 576                comments=klass._COMMENTS,
 577                var_single_tokens=klass.VAR_SINGLE_TOKENS,
 578                commands={_TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMANDS},
 579                command_prefix_tokens={
 580                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMAND_PREFIX_TOKENS
 581                },
 582                heredoc_tag_is_identifier=klass.HEREDOC_TAG_IS_IDENTIFIER,
 583                string_escapes_allowed_in_raw_strings=klass.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS,
 584                nested_comments=klass.NESTED_COMMENTS,
 585                hint_start=klass.HINT_START,
 586                tokens_preceding_hint={
 587                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.TOKENS_PRECEDING_HINT
 588                },
 589            )
 590            token_types = RsTokenTypeSettings(
 591                bit_string=_TOKEN_TYPE_TO_INDEX[TokenType.BIT_STRING],
 592                break_=_TOKEN_TYPE_TO_INDEX[TokenType.BREAK],
 593                dcolon=_TOKEN_TYPE_TO_INDEX[TokenType.DCOLON],
 594                heredoc_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEREDOC_STRING],
 595                raw_string=_TOKEN_TYPE_TO_INDEX[TokenType.RAW_STRING],
 596                hex_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEX_STRING],
 597                identifier=_TOKEN_TYPE_TO_INDEX[TokenType.IDENTIFIER],
 598                number=_TOKEN_TYPE_TO_INDEX[TokenType.NUMBER],
 599                parameter=_TOKEN_TYPE_TO_INDEX[TokenType.PARAMETER],
 600                semicolon=_TOKEN_TYPE_TO_INDEX[TokenType.SEMICOLON],
 601                string=_TOKEN_TYPE_TO_INDEX[TokenType.STRING],
 602                var=_TOKEN_TYPE_TO_INDEX[TokenType.VAR],
 603                heredoc_string_alternative=_TOKEN_TYPE_TO_INDEX[klass.HEREDOC_STRING_ALTERNATIVE],
 604                hint=_TOKEN_TYPE_TO_INDEX[TokenType.HINT],
 605            )
 606            klass._RS_TOKENIZER = RsTokenizer(settings, token_types)
 607        else:
 608            klass._RS_TOKENIZER = None
 609
 610        return klass
 611
 612
 613class Tokenizer(metaclass=_Tokenizer):
 614    SINGLE_TOKENS = {
 615        "(": TokenType.L_PAREN,
 616        ")": TokenType.R_PAREN,
 617        "[": TokenType.L_BRACKET,
 618        "]": TokenType.R_BRACKET,
 619        "{": TokenType.L_BRACE,
 620        "}": TokenType.R_BRACE,
 621        "&": TokenType.AMP,
 622        "^": TokenType.CARET,
 623        ":": TokenType.COLON,
 624        ",": TokenType.COMMA,
 625        ".": TokenType.DOT,
 626        "-": TokenType.DASH,
 627        "=": TokenType.EQ,
 628        ">": TokenType.GT,
 629        "<": TokenType.LT,
 630        "%": TokenType.MOD,
 631        "!": TokenType.NOT,
 632        "|": TokenType.PIPE,
 633        "+": TokenType.PLUS,
 634        ";": TokenType.SEMICOLON,
 635        "/": TokenType.SLASH,
 636        "\\": TokenType.BACKSLASH,
 637        "*": TokenType.STAR,
 638        "~": TokenType.TILDA,
 639        "?": TokenType.PLACEHOLDER,
 640        "@": TokenType.PARAMETER,
 641        "#": TokenType.HASH,
 642        # Used for breaking a var like x'y' but nothing else the token type doesn't matter
 643        "'": TokenType.UNKNOWN,
 644        "`": TokenType.UNKNOWN,
 645        '"': TokenType.UNKNOWN,
 646    }
 647
 648    BIT_STRINGS: t.List[str | t.Tuple[str, str]] = []
 649    BYTE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 650    HEX_STRINGS: t.List[str | t.Tuple[str, str]] = []
 651    RAW_STRINGS: t.List[str | t.Tuple[str, str]] = []
 652    HEREDOC_STRINGS: t.List[str | t.Tuple[str, str]] = []
 653    UNICODE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 654    IDENTIFIERS: t.List[str | t.Tuple[str, str]] = ['"']
 655    QUOTES: t.List[t.Tuple[str, str] | str] = ["'"]
 656    STRING_ESCAPES = ["'"]
 657    VAR_SINGLE_TOKENS: t.Set[str] = set()
 658
 659    # The strings in this list can always be used as escapes, regardless of the surrounding
 660    # identifier delimiters. By default, the closing delimiter is assumed to also act as an
 661    # identifier escape, e.g. if we use double-quotes, then they also act as escapes: "x"""
 662    IDENTIFIER_ESCAPES: t.List[str] = []
 663
 664    # Whether the heredoc tags follow the same lexical rules as unquoted identifiers
 665    HEREDOC_TAG_IS_IDENTIFIER = False
 666
 667    # Token that we'll generate as a fallback if the heredoc prefix doesn't correspond to a heredoc
 668    HEREDOC_STRING_ALTERNATIVE = TokenType.VAR
 669
 670    # Whether string escape characters function as such when placed within raw strings
 671    STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS = True
 672
 673    NESTED_COMMENTS = True
 674
 675    HINT_START = "/*+"
 676
 677    TOKENS_PRECEDING_HINT = {TokenType.SELECT, TokenType.INSERT, TokenType.UPDATE, TokenType.DELETE}
 678
 679    # Autofilled
 680    _COMMENTS: t.Dict[str, str] = {}
 681    _FORMAT_STRINGS: t.Dict[str, t.Tuple[str, TokenType]] = {}
 682    _IDENTIFIERS: t.Dict[str, str] = {}
 683    _IDENTIFIER_ESCAPES: t.Set[str] = set()
 684    _QUOTES: t.Dict[str, str] = {}
 685    _STRING_ESCAPES: t.Set[str] = set()
 686    _KEYWORD_TRIE: t.Dict = {}
 687    _RS_TOKENIZER: t.Optional[t.Any] = None
 688
 689    KEYWORDS: t.Dict[str, TokenType] = {
 690        **{f"{{%{postfix}": TokenType.BLOCK_START for postfix in ("", "+", "-")},
 691        **{f"{prefix}%}}": TokenType.BLOCK_END for prefix in ("", "+", "-")},
 692        **{f"{{{{{postfix}": TokenType.BLOCK_START for postfix in ("+", "-")},
 693        **{f"{prefix}}}}}": TokenType.BLOCK_END for prefix in ("+", "-")},
 694        HINT_START: TokenType.HINT,
 695        "==": TokenType.EQ,
 696        "::": TokenType.DCOLON,
 697        "||": TokenType.DPIPE,
 698        "|>": TokenType.PIPE_GT,
 699        ">=": TokenType.GTE,
 700        "<=": TokenType.LTE,
 701        "<>": TokenType.NEQ,
 702        "!=": TokenType.NEQ,
 703        ":=": TokenType.COLON_EQ,
 704        "<=>": TokenType.NULLSAFE_EQ,
 705        "->": TokenType.ARROW,
 706        "->>": TokenType.DARROW,
 707        "=>": TokenType.FARROW,
 708        "#>": TokenType.HASH_ARROW,
 709        "#>>": TokenType.DHASH_ARROW,
 710        "<->": TokenType.LR_ARROW,
 711        "&&": TokenType.DAMP,
 712        "??": TokenType.DQMARK,
 713        "~~~": TokenType.GLOB,
 714        "~~": TokenType.LIKE,
 715        "~~*": TokenType.ILIKE,
 716        "~*": TokenType.IRLIKE,
 717        "ALL": TokenType.ALL,
 718        "AND": TokenType.AND,
 719        "ANTI": TokenType.ANTI,
 720        "ANY": TokenType.ANY,
 721        "ASC": TokenType.ASC,
 722        "AS": TokenType.ALIAS,
 723        "ASOF": TokenType.ASOF,
 724        "AUTOINCREMENT": TokenType.AUTO_INCREMENT,
 725        "AUTO_INCREMENT": TokenType.AUTO_INCREMENT,
 726        "BEGIN": TokenType.BEGIN,
 727        "BETWEEN": TokenType.BETWEEN,
 728        "CACHE": TokenType.CACHE,
 729        "UNCACHE": TokenType.UNCACHE,
 730        "CASE": TokenType.CASE,
 731        "CHARACTER SET": TokenType.CHARACTER_SET,
 732        "CLUSTER BY": TokenType.CLUSTER_BY,
 733        "COLLATE": TokenType.COLLATE,
 734        "COLUMN": TokenType.COLUMN,
 735        "COMMIT": TokenType.COMMIT,
 736        "CONNECT BY": TokenType.CONNECT_BY,
 737        "CONSTRAINT": TokenType.CONSTRAINT,
 738        "COPY": TokenType.COPY,
 739        "CREATE": TokenType.CREATE,
 740        "CROSS": TokenType.CROSS,
 741        "CUBE": TokenType.CUBE,
 742        "CURRENT_DATE": TokenType.CURRENT_DATE,
 743        "CURRENT_SCHEMA": TokenType.CURRENT_SCHEMA,
 744        "CURRENT_TIME": TokenType.CURRENT_TIME,
 745        "CURRENT_TIMESTAMP": TokenType.CURRENT_TIMESTAMP,
 746        "CURRENT_USER": TokenType.CURRENT_USER,
 747        "DATABASE": TokenType.DATABASE,
 748        "DEFAULT": TokenType.DEFAULT,
 749        "DELETE": TokenType.DELETE,
 750        "DESC": TokenType.DESC,
 751        "DESCRIBE": TokenType.DESCRIBE,
 752        "DISTINCT": TokenType.DISTINCT,
 753        "DISTRIBUTE BY": TokenType.DISTRIBUTE_BY,
 754        "DIV": TokenType.DIV,
 755        "DROP": TokenType.DROP,
 756        "ELSE": TokenType.ELSE,
 757        "END": TokenType.END,
 758        "ENUM": TokenType.ENUM,
 759        "ESCAPE": TokenType.ESCAPE,
 760        "EXCEPT": TokenType.EXCEPT,
 761        "EXECUTE": TokenType.EXECUTE,
 762        "EXISTS": TokenType.EXISTS,
 763        "FALSE": TokenType.FALSE,
 764        "FETCH": TokenType.FETCH,
 765        "FILTER": TokenType.FILTER,
 766        "FIRST": TokenType.FIRST,
 767        "FULL": TokenType.FULL,
 768        "FUNCTION": TokenType.FUNCTION,
 769        "FOR": TokenType.FOR,
 770        "FOREIGN KEY": TokenType.FOREIGN_KEY,
 771        "FORMAT": TokenType.FORMAT,
 772        "FROM": TokenType.FROM,
 773        "GEOGRAPHY": TokenType.GEOGRAPHY,
 774        "GEOMETRY": TokenType.GEOMETRY,
 775        "GLOB": TokenType.GLOB,
 776        "GROUP BY": TokenType.GROUP_BY,
 777        "GROUPING SETS": TokenType.GROUPING_SETS,
 778        "HAVING": TokenType.HAVING,
 779        "ILIKE": TokenType.ILIKE,
 780        "IN": TokenType.IN,
 781        "INDEX": TokenType.INDEX,
 782        "INET": TokenType.INET,
 783        "INNER": TokenType.INNER,
 784        "INSERT": TokenType.INSERT,
 785        "INTERVAL": TokenType.INTERVAL,
 786        "INTERSECT": TokenType.INTERSECT,
 787        "INTO": TokenType.INTO,
 788        "IS": TokenType.IS,
 789        "ISNULL": TokenType.ISNULL,
 790        "JOIN": TokenType.JOIN,
 791        "KEEP": TokenType.KEEP,
 792        "KILL": TokenType.KILL,
 793        "LATERAL": TokenType.LATERAL,
 794        "LEFT": TokenType.LEFT,
 795        "LIKE": TokenType.LIKE,
 796        "LIMIT": TokenType.LIMIT,
 797        "LOAD": TokenType.LOAD,
 798        "LOCK": TokenType.LOCK,
 799        "MERGE": TokenType.MERGE,
 800        "NAMESPACE": TokenType.NAMESPACE,
 801        "NATURAL": TokenType.NATURAL,
 802        "NEXT": TokenType.NEXT,
 803        "NOT": TokenType.NOT,
 804        "NOTNULL": TokenType.NOTNULL,
 805        "NULL": TokenType.NULL,
 806        "OBJECT": TokenType.OBJECT,
 807        "OFFSET": TokenType.OFFSET,
 808        "ON": TokenType.ON,
 809        "OR": TokenType.OR,
 810        "XOR": TokenType.XOR,
 811        "ORDER BY": TokenType.ORDER_BY,
 812        "ORDINALITY": TokenType.ORDINALITY,
 813        "OUTER": TokenType.OUTER,
 814        "OVER": TokenType.OVER,
 815        "OVERLAPS": TokenType.OVERLAPS,
 816        "OVERWRITE": TokenType.OVERWRITE,
 817        "PARTITION": TokenType.PARTITION,
 818        "PARTITION BY": TokenType.PARTITION_BY,
 819        "PARTITIONED BY": TokenType.PARTITION_BY,
 820        "PARTITIONED_BY": TokenType.PARTITION_BY,
 821        "PERCENT": TokenType.PERCENT,
 822        "PIVOT": TokenType.PIVOT,
 823        "PRAGMA": TokenType.PRAGMA,
 824        "PRIMARY KEY": TokenType.PRIMARY_KEY,
 825        "PROCEDURE": TokenType.PROCEDURE,
 826        "QUALIFY": TokenType.QUALIFY,
 827        "RANGE": TokenType.RANGE,
 828        "RECURSIVE": TokenType.RECURSIVE,
 829        "REGEXP": TokenType.RLIKE,
 830        "RENAME": TokenType.RENAME,
 831        "REPLACE": TokenType.REPLACE,
 832        "RETURNING": TokenType.RETURNING,
 833        "REFERENCES": TokenType.REFERENCES,
 834        "RIGHT": TokenType.RIGHT,
 835        "RLIKE": TokenType.RLIKE,
 836        "ROLLBACK": TokenType.ROLLBACK,
 837        "ROLLUP": TokenType.ROLLUP,
 838        "ROW": TokenType.ROW,
 839        "ROWS": TokenType.ROWS,
 840        "SCHEMA": TokenType.SCHEMA,
 841        "SELECT": TokenType.SELECT,
 842        "SEMI": TokenType.SEMI,
 843        "SESSION": TokenType.SESSION,
 844        "SET": TokenType.SET,
 845        "SETTINGS": TokenType.SETTINGS,
 846        "SHOW": TokenType.SHOW,
 847        "SIMILAR TO": TokenType.SIMILAR_TO,
 848        "SOME": TokenType.SOME,
 849        "SORT BY": TokenType.SORT_BY,
 850        "START WITH": TokenType.START_WITH,
 851        "STRAIGHT_JOIN": TokenType.STRAIGHT_JOIN,
 852        "TABLE": TokenType.TABLE,
 853        "TABLESAMPLE": TokenType.TABLE_SAMPLE,
 854        "TEMP": TokenType.TEMPORARY,
 855        "TEMPORARY": TokenType.TEMPORARY,
 856        "THEN": TokenType.THEN,
 857        "TRUE": TokenType.TRUE,
 858        "TRUNCATE": TokenType.TRUNCATE,
 859        "UNION": TokenType.UNION,
 860        "UNKNOWN": TokenType.UNKNOWN,
 861        "UNNEST": TokenType.UNNEST,
 862        "UNPIVOT": TokenType.UNPIVOT,
 863        "UPDATE": TokenType.UPDATE,
 864        "USE": TokenType.USE,
 865        "USING": TokenType.USING,
 866        "UUID": TokenType.UUID,
 867        "VALUES": TokenType.VALUES,
 868        "VIEW": TokenType.VIEW,
 869        "VOLATILE": TokenType.VOLATILE,
 870        "WHEN": TokenType.WHEN,
 871        "WHERE": TokenType.WHERE,
 872        "WINDOW": TokenType.WINDOW,
 873        "WITH": TokenType.WITH,
 874        "APPLY": TokenType.APPLY,
 875        "ARRAY": TokenType.ARRAY,
 876        "BIT": TokenType.BIT,
 877        "BOOL": TokenType.BOOLEAN,
 878        "BOOLEAN": TokenType.BOOLEAN,
 879        "BYTE": TokenType.TINYINT,
 880        "MEDIUMINT": TokenType.MEDIUMINT,
 881        "INT1": TokenType.TINYINT,
 882        "TINYINT": TokenType.TINYINT,
 883        "INT16": TokenType.SMALLINT,
 884        "SHORT": TokenType.SMALLINT,
 885        "SMALLINT": TokenType.SMALLINT,
 886        "HUGEINT": TokenType.INT128,
 887        "UHUGEINT": TokenType.UINT128,
 888        "INT2": TokenType.SMALLINT,
 889        "INTEGER": TokenType.INT,
 890        "INT": TokenType.INT,
 891        "INT4": TokenType.INT,
 892        "INT32": TokenType.INT,
 893        "INT64": TokenType.BIGINT,
 894        "INT128": TokenType.INT128,
 895        "INT256": TokenType.INT256,
 896        "LONG": TokenType.BIGINT,
 897        "BIGINT": TokenType.BIGINT,
 898        "INT8": TokenType.TINYINT,
 899        "UINT": TokenType.UINT,
 900        "UINT128": TokenType.UINT128,
 901        "UINT256": TokenType.UINT256,
 902        "DEC": TokenType.DECIMAL,
 903        "DECIMAL": TokenType.DECIMAL,
 904        "DECIMAL32": TokenType.DECIMAL32,
 905        "DECIMAL64": TokenType.DECIMAL64,
 906        "DECIMAL128": TokenType.DECIMAL128,
 907        "DECIMAL256": TokenType.DECIMAL256,
 908        "BIGDECIMAL": TokenType.BIGDECIMAL,
 909        "BIGNUMERIC": TokenType.BIGDECIMAL,
 910        "LIST": TokenType.LIST,
 911        "MAP": TokenType.MAP,
 912        "NULLABLE": TokenType.NULLABLE,
 913        "NUMBER": TokenType.DECIMAL,
 914        "NUMERIC": TokenType.DECIMAL,
 915        "FIXED": TokenType.DECIMAL,
 916        "REAL": TokenType.FLOAT,
 917        "FLOAT": TokenType.FLOAT,
 918        "FLOAT4": TokenType.FLOAT,
 919        "FLOAT8": TokenType.DOUBLE,
 920        "DOUBLE": TokenType.DOUBLE,
 921        "DOUBLE PRECISION": TokenType.DOUBLE,
 922        "JSON": TokenType.JSON,
 923        "JSONB": TokenType.JSONB,
 924        "CHAR": TokenType.CHAR,
 925        "CHARACTER": TokenType.CHAR,
 926        "CHAR VARYING": TokenType.VARCHAR,
 927        "CHARACTER VARYING": TokenType.VARCHAR,
 928        "NCHAR": TokenType.NCHAR,
 929        "VARCHAR": TokenType.VARCHAR,
 930        "VARCHAR2": TokenType.VARCHAR,
 931        "NVARCHAR": TokenType.NVARCHAR,
 932        "NVARCHAR2": TokenType.NVARCHAR,
 933        "BPCHAR": TokenType.BPCHAR,
 934        "STR": TokenType.TEXT,
 935        "STRING": TokenType.TEXT,
 936        "TEXT": TokenType.TEXT,
 937        "LONGTEXT": TokenType.LONGTEXT,
 938        "MEDIUMTEXT": TokenType.MEDIUMTEXT,
 939        "TINYTEXT": TokenType.TINYTEXT,
 940        "CLOB": TokenType.TEXT,
 941        "LONGVARCHAR": TokenType.TEXT,
 942        "BINARY": TokenType.BINARY,
 943        "BLOB": TokenType.VARBINARY,
 944        "LONGBLOB": TokenType.LONGBLOB,
 945        "MEDIUMBLOB": TokenType.MEDIUMBLOB,
 946        "TINYBLOB": TokenType.TINYBLOB,
 947        "BYTEA": TokenType.VARBINARY,
 948        "VARBINARY": TokenType.VARBINARY,
 949        "TIME": TokenType.TIME,
 950        "TIMETZ": TokenType.TIMETZ,
 951        "TIMESTAMP": TokenType.TIMESTAMP,
 952        "TIMESTAMPTZ": TokenType.TIMESTAMPTZ,
 953        "TIMESTAMPLTZ": TokenType.TIMESTAMPLTZ,
 954        "TIMESTAMP_LTZ": TokenType.TIMESTAMPLTZ,
 955        "TIMESTAMPNTZ": TokenType.TIMESTAMPNTZ,
 956        "TIMESTAMP_NTZ": TokenType.TIMESTAMPNTZ,
 957        "DATE": TokenType.DATE,
 958        "DATETIME": TokenType.DATETIME,
 959        "INT4RANGE": TokenType.INT4RANGE,
 960        "INT4MULTIRANGE": TokenType.INT4MULTIRANGE,
 961        "INT8RANGE": TokenType.INT8RANGE,
 962        "INT8MULTIRANGE": TokenType.INT8MULTIRANGE,
 963        "NUMRANGE": TokenType.NUMRANGE,
 964        "NUMMULTIRANGE": TokenType.NUMMULTIRANGE,
 965        "TSRANGE": TokenType.TSRANGE,
 966        "TSMULTIRANGE": TokenType.TSMULTIRANGE,
 967        "TSTZRANGE": TokenType.TSTZRANGE,
 968        "TSTZMULTIRANGE": TokenType.TSTZMULTIRANGE,
 969        "DATERANGE": TokenType.DATERANGE,
 970        "DATEMULTIRANGE": TokenType.DATEMULTIRANGE,
 971        "UNIQUE": TokenType.UNIQUE,
 972        "VECTOR": TokenType.VECTOR,
 973        "STRUCT": TokenType.STRUCT,
 974        "SEQUENCE": TokenType.SEQUENCE,
 975        "VARIANT": TokenType.VARIANT,
 976        "ALTER": TokenType.ALTER,
 977        "ANALYZE": TokenType.ANALYZE,
 978        "CALL": TokenType.COMMAND,
 979        "COMMENT": TokenType.COMMENT,
 980        "EXPLAIN": TokenType.COMMAND,
 981        "GRANT": TokenType.GRANT,
 982        "REVOKE": TokenType.REVOKE,
 983        "OPTIMIZE": TokenType.COMMAND,
 984        "PREPARE": TokenType.COMMAND,
 985        "VACUUM": TokenType.COMMAND,
 986        "USER-DEFINED": TokenType.USERDEFINED,
 987        "FOR VERSION": TokenType.VERSION_SNAPSHOT,
 988        "FOR TIMESTAMP": TokenType.TIMESTAMP_SNAPSHOT,
 989    }
 990
 991    WHITE_SPACE: t.Dict[t.Optional[str], TokenType] = {
 992        " ": TokenType.SPACE,
 993        "\t": TokenType.SPACE,
 994        "\n": TokenType.BREAK,
 995        "\r": TokenType.BREAK,
 996    }
 997
 998    COMMANDS = {
 999        TokenType.COMMAND,
1000        TokenType.EXECUTE,
1001        TokenType.FETCH,
1002        TokenType.SHOW,
1003        TokenType.RENAME,
1004    }
1005
1006    COMMAND_PREFIX_TOKENS = {TokenType.SEMICOLON, TokenType.BEGIN}
1007
1008    # Handle numeric literals like in hive (3L = BIGINT)
1009    NUMERIC_LITERALS: t.Dict[str, str] = {}
1010
1011    COMMENTS = ["--", ("/*", "*/")]
1012
1013    __slots__ = (
1014        "sql",
1015        "size",
1016        "tokens",
1017        "dialect",
1018        "use_rs_tokenizer",
1019        "_start",
1020        "_current",
1021        "_line",
1022        "_col",
1023        "_comments",
1024        "_char",
1025        "_end",
1026        "_peek",
1027        "_prev_token_line",
1028        "_rs_dialect_settings",
1029    )
1030
1031    def __init__(
1032        self,
1033        dialect: DialectType = None,
1034        use_rs_tokenizer: t.Optional[bool] = None,
1035        **opts: t.Any,
1036    ) -> None:
1037        from sqlglot.dialects import Dialect
1038
1039        self.dialect = Dialect.get_or_raise(dialect)
1040
1041        # initialize `use_rs_tokenizer`, and allow it to be overwritten per Tokenizer instance
1042        self.use_rs_tokenizer = (
1043            use_rs_tokenizer if use_rs_tokenizer is not None else USE_RS_TOKENIZER
1044        )
1045
1046        if self.use_rs_tokenizer:
1047            self._rs_dialect_settings = RsTokenizerDialectSettings(
1048                unescaped_sequences=self.dialect.UNESCAPED_SEQUENCES,
1049                identifiers_can_start_with_digit=self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT,
1050                numbers_can_be_underscore_separated=self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED,
1051            )
1052
1053        self.reset()
1054
1055    def reset(self) -> None:
1056        self.sql = ""
1057        self.size = 0
1058        self.tokens: t.List[Token] = []
1059        self._start = 0
1060        self._current = 0
1061        self._line = 1
1062        self._col = 0
1063        self._comments: t.List[str] = []
1064
1065        self._char = ""
1066        self._end = False
1067        self._peek = ""
1068        self._prev_token_line = -1
1069
1070    def tokenize(self, sql: str) -> t.List[Token]:
1071        """Returns a list of tokens corresponding to the SQL string `sql`."""
1072        if self.use_rs_tokenizer:
1073            return self.tokenize_rs(sql)
1074
1075        self.reset()
1076        self.sql = sql
1077        self.size = len(sql)
1078
1079        try:
1080            self._scan()
1081        except Exception as e:
1082            start = max(self._current - 50, 0)
1083            end = min(self._current + 50, self.size - 1)
1084            context = self.sql[start:end]
1085            raise TokenError(f"Error tokenizing '{context}'") from e
1086
1087        return self.tokens
1088
1089    def _scan(self, until: t.Optional[t.Callable] = None) -> None:
1090        while self.size and not self._end:
1091            current = self._current
1092
1093            # Skip spaces here rather than iteratively calling advance() for performance reasons
1094            while current < self.size:
1095                char = self.sql[current]
1096
1097                if char.isspace() and (char == " " or char == "\t"):
1098                    current += 1
1099                else:
1100                    break
1101
1102            offset = current - self._current if current > self._current else 1
1103
1104            self._start = current
1105            self._advance(offset)
1106
1107            if not self._char.isspace():
1108                if self._char.isdigit():
1109                    self._scan_number()
1110                elif self._char in self._IDENTIFIERS:
1111                    self._scan_identifier(self._IDENTIFIERS[self._char])
1112                else:
1113                    self._scan_keywords()
1114
1115            if until and until():
1116                break
1117
1118        if self.tokens and self._comments:
1119            self.tokens[-1].comments.extend(self._comments)
1120
1121    def _chars(self, size: int) -> str:
1122        if size == 1:
1123            return self._char
1124
1125        start = self._current - 1
1126        end = start + size
1127
1128        return self.sql[start:end] if end <= self.size else ""
1129
1130    def _advance(self, i: int = 1, alnum: bool = False) -> None:
1131        if self.WHITE_SPACE.get(self._char) is TokenType.BREAK:
1132            # Ensures we don't count an extra line if we get a \r\n line break sequence
1133            if not (self._char == "\r" and self._peek == "\n"):
1134                self._col = i
1135                self._line += 1
1136        else:
1137            self._col += i
1138
1139        self._current += i
1140        self._end = self._current >= self.size
1141        self._char = self.sql[self._current - 1]
1142        self._peek = "" if self._end else self.sql[self._current]
1143
1144        if alnum and self._char.isalnum():
1145            # Here we use local variables instead of attributes for better performance
1146            _col = self._col
1147            _current = self._current
1148            _end = self._end
1149            _peek = self._peek
1150
1151            while _peek.isalnum():
1152                _col += 1
1153                _current += 1
1154                _end = _current >= self.size
1155                _peek = "" if _end else self.sql[_current]
1156
1157            self._col = _col
1158            self._current = _current
1159            self._end = _end
1160            self._peek = _peek
1161            self._char = self.sql[_current - 1]
1162
1163    @property
1164    def _text(self) -> str:
1165        return self.sql[self._start : self._current]
1166
1167    def _add(self, token_type: TokenType, text: t.Optional[str] = None) -> None:
1168        self._prev_token_line = self._line
1169
1170        if self._comments and token_type == TokenType.SEMICOLON and self.tokens:
1171            self.tokens[-1].comments.extend(self._comments)
1172            self._comments = []
1173
1174        self.tokens.append(
1175            Token(
1176                token_type,
1177                text=self._text if text is None else text,
1178                line=self._line,
1179                col=self._col,
1180                start=self._start,
1181                end=self._current - 1,
1182                comments=self._comments,
1183            )
1184        )
1185        self._comments = []
1186
1187        # If we have either a semicolon or a begin token before the command's token, we'll parse
1188        # whatever follows the command's token as a string
1189        if (
1190            token_type in self.COMMANDS
1191            and self._peek != ";"
1192            and (len(self.tokens) == 1 or self.tokens[-2].token_type in self.COMMAND_PREFIX_TOKENS)
1193        ):
1194            start = self._current
1195            tokens = len(self.tokens)
1196            self._scan(lambda: self._peek == ";")
1197            self.tokens = self.tokens[:tokens]
1198            text = self.sql[start : self._current].strip()
1199            if text:
1200                self._add(TokenType.STRING, text)
1201
1202    def _scan_keywords(self) -> None:
1203        size = 0
1204        word = None
1205        chars = self._text
1206        char = chars
1207        prev_space = False
1208        skip = False
1209        trie = self._KEYWORD_TRIE
1210        single_token = char in self.SINGLE_TOKENS
1211
1212        while chars:
1213            if skip:
1214                result = TrieResult.PREFIX
1215            else:
1216                result, trie = in_trie(trie, char.upper())
1217
1218            if result == TrieResult.FAILED:
1219                break
1220            if result == TrieResult.EXISTS:
1221                word = chars
1222
1223            end = self._current + size
1224            size += 1
1225
1226            if end < self.size:
1227                char = self.sql[end]
1228                single_token = single_token or char in self.SINGLE_TOKENS
1229                is_space = char.isspace()
1230
1231                if not is_space or not prev_space:
1232                    if is_space:
1233                        char = " "
1234                    chars += char
1235                    prev_space = is_space
1236                    skip = False
1237                else:
1238                    skip = True
1239            else:
1240                char = ""
1241                break
1242
1243        if word:
1244            if self._scan_string(word):
1245                return
1246            if self._scan_comment(word):
1247                return
1248            if prev_space or single_token or not char:
1249                self._advance(size - 1)
1250                word = word.upper()
1251                self._add(self.KEYWORDS[word], text=word)
1252                return
1253
1254        if self._char in self.SINGLE_TOKENS:
1255            self._add(self.SINGLE_TOKENS[self._char], text=self._char)
1256            return
1257
1258        self._scan_var()
1259
1260    def _scan_comment(self, comment_start: str) -> bool:
1261        if comment_start not in self._COMMENTS:
1262            return False
1263
1264        comment_start_line = self._line
1265        comment_start_size = len(comment_start)
1266        comment_end = self._COMMENTS[comment_start]
1267
1268        if comment_end:
1269            # Skip the comment's start delimiter
1270            self._advance(comment_start_size)
1271
1272            comment_count = 1
1273            comment_end_size = len(comment_end)
1274
1275            while not self._end:
1276                if self._chars(comment_end_size) == comment_end:
1277                    comment_count -= 1
1278                    if not comment_count:
1279                        break
1280
1281                self._advance(alnum=True)
1282
1283                # Nested comments are allowed by some dialects, e.g. databricks, duckdb, postgres
1284                if (
1285                    self.NESTED_COMMENTS
1286                    and not self._end
1287                    and self._chars(comment_end_size) == comment_start
1288                ):
1289                    self._advance(comment_start_size)
1290                    comment_count += 1
1291
1292            self._comments.append(self._text[comment_start_size : -comment_end_size + 1])
1293            self._advance(comment_end_size - 1)
1294        else:
1295            while not self._end and self.WHITE_SPACE.get(self._peek) is not TokenType.BREAK:
1296                self._advance(alnum=True)
1297            self._comments.append(self._text[comment_start_size:])
1298
1299        if (
1300            comment_start == self.HINT_START
1301            and self.tokens
1302            and self.tokens[-1].token_type in self.TOKENS_PRECEDING_HINT
1303        ):
1304            self._add(TokenType.HINT)
1305
1306        # Leading comment is attached to the succeeding token, whilst trailing comment to the preceding.
1307        # Multiple consecutive comments are preserved by appending them to the current comments list.
1308        if comment_start_line == self._prev_token_line:
1309            self.tokens[-1].comments.extend(self._comments)
1310            self._comments = []
1311            self._prev_token_line = self._line
1312
1313        return True
1314
1315    def _scan_number(self) -> None:
1316        if self._char == "0":
1317            peek = self._peek.upper()
1318            if peek == "B":
1319                return self._scan_bits() if self.BIT_STRINGS else self._add(TokenType.NUMBER)
1320            elif peek == "X":
1321                return self._scan_hex() if self.HEX_STRINGS else self._add(TokenType.NUMBER)
1322
1323        decimal = False
1324        scientific = 0
1325
1326        while True:
1327            if self._peek.isdigit():
1328                self._advance()
1329            elif self._peek == "." and not decimal:
1330                if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER:
1331                    return self._add(TokenType.NUMBER)
1332                decimal = True
1333                self._advance()
1334            elif self._peek in ("-", "+") and scientific == 1:
1335                scientific += 1
1336                self._advance()
1337            elif self._peek.upper() == "E" and not scientific:
1338                scientific += 1
1339                self._advance()
1340            elif self._peek.isidentifier():
1341                number_text = self._text
1342                literal = ""
1343
1344                while self._peek.strip() and self._peek not in self.SINGLE_TOKENS:
1345                    literal += self._peek
1346                    self._advance()
1347
1348                token_type = self.KEYWORDS.get(self.NUMERIC_LITERALS.get(literal.upper(), ""))
1349
1350                if token_type:
1351                    self._add(TokenType.NUMBER, number_text)
1352                    self._add(TokenType.DCOLON, "::")
1353                    return self._add(token_type, literal)
1354                else:
1355                    replaced = literal.replace("_", "")
1356                    if self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED and replaced.isdigit():
1357                        return self._add(TokenType.NUMBER, number_text + replaced)
1358                    if self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT:
1359                        return self._add(TokenType.VAR)
1360
1361                self._advance(-len(literal))
1362                return self._add(TokenType.NUMBER, number_text)
1363            else:
1364                return self._add(TokenType.NUMBER)
1365
1366    def _scan_bits(self) -> None:
1367        self._advance()
1368        value = self._extract_value()
1369        try:
1370            # If `value` can't be converted to a binary, fallback to tokenizing it as an identifier
1371            int(value, 2)
1372            self._add(TokenType.BIT_STRING, value[2:])  # Drop the 0b
1373        except ValueError:
1374            self._add(TokenType.IDENTIFIER)
1375
1376    def _scan_hex(self) -> None:
1377        self._advance()
1378        value = self._extract_value()
1379        try:
1380            # If `value` can't be converted to a hex, fallback to tokenizing it as an identifier
1381            int(value, 16)
1382            self._add(TokenType.HEX_STRING, value[2:])  # Drop the 0x
1383        except ValueError:
1384            self._add(TokenType.IDENTIFIER)
1385
1386    def _extract_value(self) -> str:
1387        while True:
1388            char = self._peek.strip()
1389            if char and char not in self.SINGLE_TOKENS:
1390                self._advance(alnum=True)
1391            else:
1392                break
1393
1394        return self._text
1395
1396    def _scan_string(self, start: str) -> bool:
1397        base = None
1398        token_type = TokenType.STRING
1399
1400        if start in self._QUOTES:
1401            end = self._QUOTES[start]
1402        elif start in self._FORMAT_STRINGS:
1403            end, token_type = self._FORMAT_STRINGS[start]
1404
1405            if token_type == TokenType.HEX_STRING:
1406                base = 16
1407            elif token_type == TokenType.BIT_STRING:
1408                base = 2
1409            elif token_type == TokenType.HEREDOC_STRING:
1410                self._advance()
1411
1412                if self._char == end:
1413                    tag = ""
1414                else:
1415                    tag = self._extract_string(
1416                        end,
1417                        raw_string=True,
1418                        raise_unmatched=not self.HEREDOC_TAG_IS_IDENTIFIER,
1419                    )
1420
1421                if tag and self.HEREDOC_TAG_IS_IDENTIFIER and (self._end or not tag.isidentifier()):
1422                    if not self._end:
1423                        self._advance(-1)
1424
1425                    self._advance(-len(tag))
1426                    self._add(self.HEREDOC_STRING_ALTERNATIVE)
1427                    return True
1428
1429                end = f"{start}{tag}{end}"
1430        else:
1431            return False
1432
1433        self._advance(len(start))
1434        text = self._extract_string(end, raw_string=token_type == TokenType.RAW_STRING)
1435
1436        if base and text:
1437            try:
1438                int(text, base)
1439            except Exception:
1440                raise TokenError(
1441                    f"Numeric string contains invalid characters from {self._line}:{self._start}"
1442                )
1443
1444        self._add(token_type, text)
1445        return True
1446
1447    def _scan_identifier(self, identifier_end: str) -> None:
1448        self._advance()
1449        text = self._extract_string(
1450            identifier_end, escapes=self._IDENTIFIER_ESCAPES | {identifier_end}
1451        )
1452        self._add(TokenType.IDENTIFIER, text)
1453
1454    def _scan_var(self) -> None:
1455        while True:
1456            char = self._peek.strip()
1457            if char and (char in self.VAR_SINGLE_TOKENS or char not in self.SINGLE_TOKENS):
1458                self._advance(alnum=True)
1459            else:
1460                break
1461
1462        self._add(
1463            TokenType.VAR
1464            if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER
1465            else self.KEYWORDS.get(self._text.upper(), TokenType.VAR)
1466        )
1467
1468    def _extract_string(
1469        self,
1470        delimiter: str,
1471        escapes: t.Optional[t.Set[str]] = None,
1472        raw_string: bool = False,
1473        raise_unmatched: bool = True,
1474    ) -> str:
1475        text = ""
1476        delim_size = len(delimiter)
1477        escapes = self._STRING_ESCAPES if escapes is None else escapes
1478
1479        while True:
1480            if (
1481                not raw_string
1482                and self.dialect.UNESCAPED_SEQUENCES
1483                and self._peek
1484                and self._char in self.STRING_ESCAPES
1485            ):
1486                unescaped_sequence = self.dialect.UNESCAPED_SEQUENCES.get(self._char + self._peek)
1487                if unescaped_sequence:
1488                    self._advance(2)
1489                    text += unescaped_sequence
1490                    continue
1491            if (
1492                (self.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS or not raw_string)
1493                and self._char in escapes
1494                and (self._peek == delimiter or self._peek in escapes)
1495                and (self._char not in self._QUOTES or self._char == self._peek)
1496            ):
1497                if self._peek == delimiter:
1498                    text += self._peek
1499                else:
1500                    text += self._char + self._peek
1501
1502                if self._current + 1 < self.size:
1503                    self._advance(2)
1504                else:
1505                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._current}")
1506            else:
1507                if self._chars(delim_size) == delimiter:
1508                    if delim_size > 1:
1509                        self._advance(delim_size - 1)
1510                    break
1511
1512                if self._end:
1513                    if not raise_unmatched:
1514                        return text + self._char
1515
1516                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._start}")
1517
1518                current = self._current - 1
1519                self._advance(alnum=True)
1520                text += self.sql[current : self._current - 1]
1521
1522        return text
1523
1524    def tokenize_rs(self, sql: str) -> t.List[Token]:
1525        if not self._RS_TOKENIZER:
1526            raise SqlglotError("Rust tokenizer is not available")
1527
1528        tokens, error_msg = self._RS_TOKENIZER.tokenize(sql, self._rs_dialect_settings)
1529        for token in tokens:
1530            token.token_type = _ALL_TOKEN_TYPES[token.token_type_index]
1531
1532        # Setting this here so partial token lists can be inspected even if there is a failure
1533        self.tokens = tokens
1534
1535        if error_msg is not None:
1536            raise TokenError(error_msg)
1537
1538        return tokens