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    EXCLAMATION = auto()
  92
  93    URI_START = auto()
  94
  95    BLOCK_START = auto()
  96    BLOCK_END = auto()
  97
  98    SPACE = auto()
  99    BREAK = auto()
 100
 101    STRING = auto()
 102    NUMBER = auto()
 103    IDENTIFIER = auto()
 104    DATABASE = auto()
 105    COLUMN = auto()
 106    COLUMN_DEF = auto()
 107    SCHEMA = auto()
 108    TABLE = auto()
 109    WAREHOUSE = auto()
 110    STAGE = auto()
 111    STREAMLIT = auto()
 112    VAR = auto()
 113    BIT_STRING = auto()
 114    HEX_STRING = auto()
 115    BYTE_STRING = auto()
 116    NATIONAL_STRING = auto()
 117    RAW_STRING = auto()
 118    HEREDOC_STRING = auto()
 119    UNICODE_STRING = auto()
 120
 121    # types
 122    BIT = auto()
 123    BOOLEAN = auto()
 124    TINYINT = auto()
 125    UTINYINT = auto()
 126    SMALLINT = auto()
 127    USMALLINT = auto()
 128    MEDIUMINT = auto()
 129    UMEDIUMINT = auto()
 130    INT = auto()
 131    UINT = auto()
 132    BIGINT = auto()
 133    UBIGINT = auto()
 134    INT128 = auto()
 135    UINT128 = auto()
 136    INT256 = auto()
 137    UINT256 = auto()
 138    FLOAT = auto()
 139    DOUBLE = auto()
 140    UDOUBLE = auto()
 141    DECIMAL = auto()
 142    DECIMAL32 = auto()
 143    DECIMAL64 = auto()
 144    DECIMAL128 = auto()
 145    DECIMAL256 = auto()
 146    UDECIMAL = auto()
 147    BIGDECIMAL = auto()
 148    CHAR = auto()
 149    NCHAR = auto()
 150    VARCHAR = auto()
 151    NVARCHAR = auto()
 152    BPCHAR = auto()
 153    TEXT = auto()
 154    MEDIUMTEXT = auto()
 155    LONGTEXT = auto()
 156    BLOB = auto()
 157    MEDIUMBLOB = auto()
 158    LONGBLOB = auto()
 159    TINYBLOB = auto()
 160    TINYTEXT = auto()
 161    NAME = auto()
 162    BINARY = auto()
 163    VARBINARY = auto()
 164    JSON = auto()
 165    JSONB = auto()
 166    TIME = auto()
 167    TIMETZ = auto()
 168    TIMESTAMP = auto()
 169    TIMESTAMPTZ = auto()
 170    TIMESTAMPLTZ = auto()
 171    TIMESTAMPNTZ = auto()
 172    TIMESTAMP_S = auto()
 173    TIMESTAMP_MS = auto()
 174    TIMESTAMP_NS = auto()
 175    DATETIME = auto()
 176    DATETIME2 = auto()
 177    DATETIME64 = auto()
 178    SMALLDATETIME = auto()
 179    DATE = auto()
 180    DATE32 = auto()
 181    INT4RANGE = auto()
 182    INT4MULTIRANGE = auto()
 183    INT8RANGE = auto()
 184    INT8MULTIRANGE = auto()
 185    NUMRANGE = auto()
 186    NUMMULTIRANGE = auto()
 187    TSRANGE = auto()
 188    TSMULTIRANGE = auto()
 189    TSTZRANGE = auto()
 190    TSTZMULTIRANGE = auto()
 191    DATERANGE = auto()
 192    DATEMULTIRANGE = auto()
 193    UUID = auto()
 194    GEOGRAPHY = auto()
 195    GEOGRAPHYPOINT = auto()
 196    NULLABLE = auto()
 197    GEOMETRY = auto()
 198    POINT = auto()
 199    RING = auto()
 200    LINESTRING = auto()
 201    MULTILINESTRING = auto()
 202    POLYGON = auto()
 203    MULTIPOLYGON = auto()
 204    HLLSKETCH = auto()
 205    HSTORE = auto()
 206    SUPER = auto()
 207    SERIAL = auto()
 208    SMALLSERIAL = auto()
 209    BIGSERIAL = auto()
 210    XML = auto()
 211    YEAR = auto()
 212    USERDEFINED = auto()
 213    MONEY = auto()
 214    SMALLMONEY = auto()
 215    ROWVERSION = auto()
 216    IMAGE = auto()
 217    VARIANT = auto()
 218    OBJECT = auto()
 219    INET = auto()
 220    IPADDRESS = auto()
 221    IPPREFIX = auto()
 222    IPV4 = auto()
 223    IPV6 = auto()
 224    ENUM = auto()
 225    ENUM8 = auto()
 226    ENUM16 = auto()
 227    FIXEDSTRING = auto()
 228    LOWCARDINALITY = auto()
 229    NESTED = auto()
 230    AGGREGATEFUNCTION = auto()
 231    SIMPLEAGGREGATEFUNCTION = auto()
 232    TDIGEST = auto()
 233    UNKNOWN = auto()
 234    VECTOR = auto()
 235    DYNAMIC = auto()
 236    VOID = auto()
 237
 238    # keywords
 239    ALIAS = auto()
 240    ALTER = auto()
 241    ALL = auto()
 242    ANTI = auto()
 243    ANY = auto()
 244    APPLY = auto()
 245    ARRAY = auto()
 246    ASC = auto()
 247    ASOF = auto()
 248    ATTACH = auto()
 249    AUTO_INCREMENT = auto()
 250    BEGIN = auto()
 251    BETWEEN = auto()
 252    BULK_COLLECT_INTO = auto()
 253    CACHE = auto()
 254    CASE = auto()
 255    CHARACTER_SET = auto()
 256    CLUSTER_BY = auto()
 257    COLLATE = auto()
 258    COMMAND = auto()
 259    COMMENT = auto()
 260    COMMIT = auto()
 261    CONNECT_BY = auto()
 262    CONSTRAINT = auto()
 263    COPY = auto()
 264    CREATE = auto()
 265    CROSS = auto()
 266    CUBE = auto()
 267    CURRENT_DATE = auto()
 268    CURRENT_DATETIME = auto()
 269    CURRENT_SCHEMA = auto()
 270    CURRENT_TIME = auto()
 271    CURRENT_TIMESTAMP = auto()
 272    CURRENT_USER = auto()
 273    DECLARE = auto()
 274    DEFAULT = auto()
 275    DELETE = auto()
 276    DESC = auto()
 277    DESCRIBE = auto()
 278    DETACH = auto()
 279    DICTIONARY = auto()
 280    DISTINCT = auto()
 281    DISTRIBUTE_BY = auto()
 282    DIV = auto()
 283    DROP = auto()
 284    ELSE = auto()
 285    END = auto()
 286    ESCAPE = auto()
 287    EXCEPT = auto()
 288    EXECUTE = auto()
 289    EXISTS = auto()
 290    FALSE = auto()
 291    FETCH = auto()
 292    FILE_FORMAT = auto()
 293    FILTER = auto()
 294    FINAL = auto()
 295    FIRST = auto()
 296    FOR = auto()
 297    FORCE = auto()
 298    FOREIGN_KEY = auto()
 299    FORMAT = auto()
 300    FROM = auto()
 301    FULL = auto()
 302    FUNCTION = auto()
 303    GET = auto()
 304    GLOB = auto()
 305    GLOBAL = auto()
 306    GRANT = auto()
 307    GROUP_BY = auto()
 308    GROUPING_SETS = auto()
 309    HAVING = auto()
 310    HINT = auto()
 311    IGNORE = auto()
 312    ILIKE = auto()
 313    IN = auto()
 314    INDEX = auto()
 315    INNER = auto()
 316    INSERT = auto()
 317    INSTALL = auto()
 318    INTERSECT = auto()
 319    INTERVAL = auto()
 320    INTO = auto()
 321    INTRODUCER = auto()
 322    IRLIKE = auto()
 323    IS = auto()
 324    ISNULL = auto()
 325    JOIN = auto()
 326    JOIN_MARKER = auto()
 327    KEEP = auto()
 328    KEY = auto()
 329    KILL = auto()
 330    LANGUAGE = auto()
 331    LATERAL = auto()
 332    LEFT = auto()
 333    LIKE = auto()
 334    LIMIT = auto()
 335    LIST = auto()
 336    LOAD = auto()
 337    LOCK = auto()
 338    MAP = auto()
 339    MATCH_CONDITION = auto()
 340    MATCH_RECOGNIZE = auto()
 341    MEMBER_OF = auto()
 342    MERGE = auto()
 343    MOD = auto()
 344    MODEL = auto()
 345    NATURAL = auto()
 346    NEXT = auto()
 347    NOTHING = auto()
 348    NOTNULL = auto()
 349    NULL = auto()
 350    OBJECT_IDENTIFIER = auto()
 351    OFFSET = auto()
 352    ON = auto()
 353    ONLY = auto()
 354    OPERATOR = auto()
 355    ORDER_BY = auto()
 356    ORDER_SIBLINGS_BY = auto()
 357    ORDERED = auto()
 358    ORDINALITY = auto()
 359    OUTER = auto()
 360    OVER = auto()
 361    OVERLAPS = auto()
 362    OVERWRITE = auto()
 363    PARTITION = auto()
 364    PARTITION_BY = auto()
 365    PERCENT = auto()
 366    PIVOT = auto()
 367    PLACEHOLDER = auto()
 368    POSITIONAL = auto()
 369    PRAGMA = auto()
 370    PREWHERE = auto()
 371    PRIMARY_KEY = auto()
 372    PROCEDURE = auto()
 373    PROPERTIES = auto()
 374    PSEUDO_TYPE = auto()
 375    PUT = auto()
 376    QUALIFY = auto()
 377    QUOTE = auto()
 378    RANGE = auto()
 379    RECURSIVE = auto()
 380    REFRESH = auto()
 381    RENAME = auto()
 382    REPLACE = auto()
 383    RETURNING = auto()
 384    REVOKE = auto()
 385    REFERENCES = auto()
 386    RIGHT = auto()
 387    RLIKE = auto()
 388    ROLLBACK = auto()
 389    ROLLUP = auto()
 390    ROW = auto()
 391    ROWS = auto()
 392    SELECT = auto()
 393    SEMI = auto()
 394    SEPARATOR = auto()
 395    SEQUENCE = auto()
 396    SERDE_PROPERTIES = auto()
 397    SET = auto()
 398    SETTINGS = auto()
 399    SHOW = auto()
 400    SIMILAR_TO = auto()
 401    SOME = auto()
 402    SORT_BY = auto()
 403    SOUNDS_LIKE = auto()
 404    START_WITH = auto()
 405    STORAGE_INTEGRATION = auto()
 406    STRAIGHT_JOIN = auto()
 407    STRUCT = auto()
 408    SUMMARIZE = auto()
 409    TABLE_SAMPLE = auto()
 410    TAG = auto()
 411    TEMPORARY = auto()
 412    TOP = auto()
 413    THEN = auto()
 414    TRUE = auto()
 415    TRUNCATE = auto()
 416    UNCACHE = auto()
 417    UNION = auto()
 418    UNNEST = auto()
 419    UNPIVOT = auto()
 420    UPDATE = auto()
 421    USE = auto()
 422    USING = auto()
 423    VALUES = auto()
 424    VIEW = auto()
 425    SEMANTIC_VIEW = auto()
 426    VOLATILE = auto()
 427    WHEN = auto()
 428    WHERE = auto()
 429    WINDOW = auto()
 430    WITH = auto()
 431    UNIQUE = auto()
 432    UTC_DATE = auto()
 433    UTC_TIME = auto()
 434    UTC_TIMESTAMP = auto()
 435    VERSION_SNAPSHOT = auto()
 436    TIMESTAMP_SNAPSHOT = auto()
 437    OPTION = auto()
 438    SINK = auto()
 439    SOURCE = auto()
 440    ANALYZE = auto()
 441    NAMESPACE = auto()
 442    EXPORT = auto()
 443
 444    # sentinel
 445    HIVE_TOKEN_STREAM = auto()
 446
 447
 448_ALL_TOKEN_TYPES = list(TokenType)
 449_TOKEN_TYPE_TO_INDEX = {token_type: i for i, token_type in enumerate(_ALL_TOKEN_TYPES)}
 450
 451
 452class Token:
 453    __slots__ = ("token_type", "text", "line", "col", "start", "end", "comments")
 454
 455    @classmethod
 456    def number(cls, number: int) -> Token:
 457        """Returns a NUMBER token with `number` as its text."""
 458        return cls(TokenType.NUMBER, str(number))
 459
 460    @classmethod
 461    def string(cls, string: str) -> Token:
 462        """Returns a STRING token with `string` as its text."""
 463        return cls(TokenType.STRING, string)
 464
 465    @classmethod
 466    def identifier(cls, identifier: str) -> Token:
 467        """Returns an IDENTIFIER token with `identifier` as its text."""
 468        return cls(TokenType.IDENTIFIER, identifier)
 469
 470    @classmethod
 471    def var(cls, var: str) -> Token:
 472        """Returns an VAR token with `var` as its text."""
 473        return cls(TokenType.VAR, var)
 474
 475    def __init__(
 476        self,
 477        token_type: TokenType,
 478        text: str,
 479        line: int = 1,
 480        col: int = 1,
 481        start: int = 0,
 482        end: int = 0,
 483        comments: t.Optional[t.List[str]] = None,
 484    ) -> None:
 485        """Token initializer.
 486
 487        Args:
 488            token_type: The TokenType Enum.
 489            text: The text of the token.
 490            line: The line that the token ends on.
 491            col: The column that the token ends on.
 492            start: The start index of the token.
 493            end: The ending index of the token.
 494            comments: The comments to attach to the token.
 495        """
 496        self.token_type = token_type
 497        self.text = text
 498        self.line = line
 499        self.col = col
 500        self.start = start
 501        self.end = end
 502        self.comments = [] if comments is None else comments
 503
 504    def __repr__(self) -> str:
 505        attributes = ", ".join(f"{k}: {getattr(self, k)}" for k in self.__slots__)
 506        return f"<Token {attributes}>"
 507
 508
 509class _Tokenizer(type):
 510    def __new__(cls, clsname, bases, attrs):
 511        klass = super().__new__(cls, clsname, bases, attrs)
 512
 513        def _convert_quotes(arr: t.List[str | t.Tuple[str, str]]) -> t.Dict[str, str]:
 514            return dict(
 515                (item, item) if isinstance(item, str) else (item[0], item[1]) for item in arr
 516            )
 517
 518        def _quotes_to_format(
 519            token_type: TokenType, arr: t.List[str | t.Tuple[str, str]]
 520        ) -> t.Dict[str, t.Tuple[str, TokenType]]:
 521            return {k: (v, token_type) for k, v in _convert_quotes(arr).items()}
 522
 523        klass._QUOTES = _convert_quotes(klass.QUOTES)
 524        klass._IDENTIFIERS = _convert_quotes(klass.IDENTIFIERS)
 525
 526        klass._FORMAT_STRINGS = {
 527            **{
 528                p + s: (e, TokenType.NATIONAL_STRING)
 529                for s, e in klass._QUOTES.items()
 530                for p in ("n", "N")
 531            },
 532            **_quotes_to_format(TokenType.BIT_STRING, klass.BIT_STRINGS),
 533            **_quotes_to_format(TokenType.BYTE_STRING, klass.BYTE_STRINGS),
 534            **_quotes_to_format(TokenType.HEX_STRING, klass.HEX_STRINGS),
 535            **_quotes_to_format(TokenType.RAW_STRING, klass.RAW_STRINGS),
 536            **_quotes_to_format(TokenType.HEREDOC_STRING, klass.HEREDOC_STRINGS),
 537            **_quotes_to_format(TokenType.UNICODE_STRING, klass.UNICODE_STRINGS),
 538        }
 539
 540        klass._STRING_ESCAPES = set(klass.STRING_ESCAPES)
 541        klass._IDENTIFIER_ESCAPES = set(klass.IDENTIFIER_ESCAPES)
 542        klass._COMMENTS = {
 543            **dict(
 544                (comment, None) if isinstance(comment, str) else (comment[0], comment[1])
 545                for comment in klass.COMMENTS
 546            ),
 547            "{#": "#}",  # Ensure Jinja comments are tokenized correctly in all dialects
 548        }
 549        if klass.HINT_START in klass.KEYWORDS:
 550            klass._COMMENTS[klass.HINT_START] = "*/"
 551
 552        klass._KEYWORD_TRIE = new_trie(
 553            key.upper()
 554            for key in (
 555                *klass.KEYWORDS,
 556                *klass._COMMENTS,
 557                *klass._QUOTES,
 558                *klass._FORMAT_STRINGS,
 559            )
 560            if " " in key or any(single in key for single in klass.SINGLE_TOKENS)
 561        )
 562
 563        if USE_RS_TOKENIZER:
 564            settings = RsTokenizerSettings(
 565                white_space={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.WHITE_SPACE.items()},
 566                single_tokens={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.SINGLE_TOKENS.items()},
 567                keywords={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.KEYWORDS.items()},
 568                numeric_literals=klass.NUMERIC_LITERALS,
 569                identifiers=klass._IDENTIFIERS,
 570                identifier_escapes=klass._IDENTIFIER_ESCAPES,
 571                string_escapes=klass._STRING_ESCAPES,
 572                quotes=klass._QUOTES,
 573                format_strings={
 574                    k: (v1, _TOKEN_TYPE_TO_INDEX[v2])
 575                    for k, (v1, v2) in klass._FORMAT_STRINGS.items()
 576                },
 577                has_bit_strings=bool(klass.BIT_STRINGS),
 578                has_hex_strings=bool(klass.HEX_STRINGS),
 579                comments=klass._COMMENTS,
 580                var_single_tokens=klass.VAR_SINGLE_TOKENS,
 581                commands={_TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMANDS},
 582                command_prefix_tokens={
 583                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMAND_PREFIX_TOKENS
 584                },
 585                heredoc_tag_is_identifier=klass.HEREDOC_TAG_IS_IDENTIFIER,
 586                string_escapes_allowed_in_raw_strings=klass.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS,
 587                nested_comments=klass.NESTED_COMMENTS,
 588                hint_start=klass.HINT_START,
 589                tokens_preceding_hint={
 590                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.TOKENS_PRECEDING_HINT
 591                },
 592            )
 593            token_types = RsTokenTypeSettings(
 594                bit_string=_TOKEN_TYPE_TO_INDEX[TokenType.BIT_STRING],
 595                break_=_TOKEN_TYPE_TO_INDEX[TokenType.BREAK],
 596                dcolon=_TOKEN_TYPE_TO_INDEX[TokenType.DCOLON],
 597                heredoc_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEREDOC_STRING],
 598                raw_string=_TOKEN_TYPE_TO_INDEX[TokenType.RAW_STRING],
 599                hex_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEX_STRING],
 600                identifier=_TOKEN_TYPE_TO_INDEX[TokenType.IDENTIFIER],
 601                number=_TOKEN_TYPE_TO_INDEX[TokenType.NUMBER],
 602                parameter=_TOKEN_TYPE_TO_INDEX[TokenType.PARAMETER],
 603                semicolon=_TOKEN_TYPE_TO_INDEX[TokenType.SEMICOLON],
 604                string=_TOKEN_TYPE_TO_INDEX[TokenType.STRING],
 605                var=_TOKEN_TYPE_TO_INDEX[TokenType.VAR],
 606                heredoc_string_alternative=_TOKEN_TYPE_TO_INDEX[klass.HEREDOC_STRING_ALTERNATIVE],
 607                hint=_TOKEN_TYPE_TO_INDEX[TokenType.HINT],
 608            )
 609            klass._RS_TOKENIZER = RsTokenizer(settings, token_types)
 610        else:
 611            klass._RS_TOKENIZER = None
 612
 613        return klass
 614
 615
 616class Tokenizer(metaclass=_Tokenizer):
 617    SINGLE_TOKENS = {
 618        "(": TokenType.L_PAREN,
 619        ")": TokenType.R_PAREN,
 620        "[": TokenType.L_BRACKET,
 621        "]": TokenType.R_BRACKET,
 622        "{": TokenType.L_BRACE,
 623        "}": TokenType.R_BRACE,
 624        "&": TokenType.AMP,
 625        "^": TokenType.CARET,
 626        ":": TokenType.COLON,
 627        ",": TokenType.COMMA,
 628        ".": TokenType.DOT,
 629        "-": TokenType.DASH,
 630        "=": TokenType.EQ,
 631        ">": TokenType.GT,
 632        "<": TokenType.LT,
 633        "%": TokenType.MOD,
 634        "!": TokenType.NOT,
 635        "|": TokenType.PIPE,
 636        "+": TokenType.PLUS,
 637        ";": TokenType.SEMICOLON,
 638        "/": TokenType.SLASH,
 639        "\\": TokenType.BACKSLASH,
 640        "*": TokenType.STAR,
 641        "~": TokenType.TILDA,
 642        "?": TokenType.PLACEHOLDER,
 643        "@": TokenType.PARAMETER,
 644        "#": TokenType.HASH,
 645        # Used for breaking a var like x'y' but nothing else the token type doesn't matter
 646        "'": TokenType.UNKNOWN,
 647        "`": TokenType.UNKNOWN,
 648        '"': TokenType.UNKNOWN,
 649    }
 650
 651    BIT_STRINGS: t.List[str | t.Tuple[str, str]] = []
 652    BYTE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 653    HEX_STRINGS: t.List[str | t.Tuple[str, str]] = []
 654    RAW_STRINGS: t.List[str | t.Tuple[str, str]] = []
 655    HEREDOC_STRINGS: t.List[str | t.Tuple[str, str]] = []
 656    UNICODE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 657    IDENTIFIERS: t.List[str | t.Tuple[str, str]] = ['"']
 658    QUOTES: t.List[t.Tuple[str, str] | str] = ["'"]
 659    STRING_ESCAPES = ["'"]
 660    VAR_SINGLE_TOKENS: t.Set[str] = set()
 661
 662    # The strings in this list can always be used as escapes, regardless of the surrounding
 663    # identifier delimiters. By default, the closing delimiter is assumed to also act as an
 664    # identifier escape, e.g. if we use double-quotes, then they also act as escapes: "x"""
 665    IDENTIFIER_ESCAPES: t.List[str] = []
 666
 667    # Whether the heredoc tags follow the same lexical rules as unquoted identifiers
 668    HEREDOC_TAG_IS_IDENTIFIER = False
 669
 670    # Token that we'll generate as a fallback if the heredoc prefix doesn't correspond to a heredoc
 671    HEREDOC_STRING_ALTERNATIVE = TokenType.VAR
 672
 673    # Whether string escape characters function as such when placed within raw strings
 674    STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS = True
 675
 676    NESTED_COMMENTS = True
 677
 678    HINT_START = "/*+"
 679
 680    TOKENS_PRECEDING_HINT = {TokenType.SELECT, TokenType.INSERT, TokenType.UPDATE, TokenType.DELETE}
 681
 682    # Autofilled
 683    _COMMENTS: t.Dict[str, str] = {}
 684    _FORMAT_STRINGS: t.Dict[str, t.Tuple[str, TokenType]] = {}
 685    _IDENTIFIERS: t.Dict[str, str] = {}
 686    _IDENTIFIER_ESCAPES: t.Set[str] = set()
 687    _QUOTES: t.Dict[str, str] = {}
 688    _STRING_ESCAPES: t.Set[str] = set()
 689    _KEYWORD_TRIE: t.Dict = {}
 690    _RS_TOKENIZER: t.Optional[t.Any] = None
 691
 692    KEYWORDS: t.Dict[str, TokenType] = {
 693        **{f"{{%{postfix}": TokenType.BLOCK_START for postfix in ("", "+", "-")},
 694        **{f"{prefix}%}}": TokenType.BLOCK_END for prefix in ("", "+", "-")},
 695        **{f"{{{{{postfix}": TokenType.BLOCK_START for postfix in ("+", "-")},
 696        **{f"{prefix}}}}}": TokenType.BLOCK_END for prefix in ("+", "-")},
 697        HINT_START: TokenType.HINT,
 698        "==": TokenType.EQ,
 699        "::": TokenType.DCOLON,
 700        "||": TokenType.DPIPE,
 701        "|>": TokenType.PIPE_GT,
 702        ">=": TokenType.GTE,
 703        "<=": TokenType.LTE,
 704        "<>": TokenType.NEQ,
 705        "!=": TokenType.NEQ,
 706        ":=": TokenType.COLON_EQ,
 707        "<=>": TokenType.NULLSAFE_EQ,
 708        "->": TokenType.ARROW,
 709        "->>": TokenType.DARROW,
 710        "=>": TokenType.FARROW,
 711        "#>": TokenType.HASH_ARROW,
 712        "#>>": TokenType.DHASH_ARROW,
 713        "<->": TokenType.LR_ARROW,
 714        "&&": TokenType.DAMP,
 715        "??": TokenType.DQMARK,
 716        "~~~": TokenType.GLOB,
 717        "~~": TokenType.LIKE,
 718        "~~*": TokenType.ILIKE,
 719        "~*": TokenType.IRLIKE,
 720        "ALL": TokenType.ALL,
 721        "AND": TokenType.AND,
 722        "ANTI": TokenType.ANTI,
 723        "ANY": TokenType.ANY,
 724        "ASC": TokenType.ASC,
 725        "AS": TokenType.ALIAS,
 726        "ASOF": TokenType.ASOF,
 727        "AUTOINCREMENT": TokenType.AUTO_INCREMENT,
 728        "AUTO_INCREMENT": TokenType.AUTO_INCREMENT,
 729        "BEGIN": TokenType.BEGIN,
 730        "BETWEEN": TokenType.BETWEEN,
 731        "CACHE": TokenType.CACHE,
 732        "UNCACHE": TokenType.UNCACHE,
 733        "CASE": TokenType.CASE,
 734        "CHARACTER SET": TokenType.CHARACTER_SET,
 735        "CLUSTER BY": TokenType.CLUSTER_BY,
 736        "COLLATE": TokenType.COLLATE,
 737        "COLUMN": TokenType.COLUMN,
 738        "COMMIT": TokenType.COMMIT,
 739        "CONNECT BY": TokenType.CONNECT_BY,
 740        "CONSTRAINT": TokenType.CONSTRAINT,
 741        "COPY": TokenType.COPY,
 742        "CREATE": TokenType.CREATE,
 743        "CROSS": TokenType.CROSS,
 744        "CUBE": TokenType.CUBE,
 745        "CURRENT_DATE": TokenType.CURRENT_DATE,
 746        "CURRENT_SCHEMA": TokenType.CURRENT_SCHEMA,
 747        "CURRENT_TIME": TokenType.CURRENT_TIME,
 748        "CURRENT_TIMESTAMP": TokenType.CURRENT_TIMESTAMP,
 749        "CURRENT_USER": TokenType.CURRENT_USER,
 750        "DATABASE": TokenType.DATABASE,
 751        "DEFAULT": TokenType.DEFAULT,
 752        "DELETE": TokenType.DELETE,
 753        "DESC": TokenType.DESC,
 754        "DESCRIBE": TokenType.DESCRIBE,
 755        "DISTINCT": TokenType.DISTINCT,
 756        "DISTRIBUTE BY": TokenType.DISTRIBUTE_BY,
 757        "DIV": TokenType.DIV,
 758        "DROP": TokenType.DROP,
 759        "ELSE": TokenType.ELSE,
 760        "END": TokenType.END,
 761        "ENUM": TokenType.ENUM,
 762        "ESCAPE": TokenType.ESCAPE,
 763        "EXCEPT": TokenType.EXCEPT,
 764        "EXECUTE": TokenType.EXECUTE,
 765        "EXISTS": TokenType.EXISTS,
 766        "FALSE": TokenType.FALSE,
 767        "FETCH": TokenType.FETCH,
 768        "FILTER": TokenType.FILTER,
 769        "FIRST": TokenType.FIRST,
 770        "FULL": TokenType.FULL,
 771        "FUNCTION": TokenType.FUNCTION,
 772        "FOR": TokenType.FOR,
 773        "FOREIGN KEY": TokenType.FOREIGN_KEY,
 774        "FORMAT": TokenType.FORMAT,
 775        "FROM": TokenType.FROM,
 776        "GEOGRAPHY": TokenType.GEOGRAPHY,
 777        "GEOMETRY": TokenType.GEOMETRY,
 778        "GLOB": TokenType.GLOB,
 779        "GROUP BY": TokenType.GROUP_BY,
 780        "GROUPING SETS": TokenType.GROUPING_SETS,
 781        "HAVING": TokenType.HAVING,
 782        "ILIKE": TokenType.ILIKE,
 783        "IN": TokenType.IN,
 784        "INDEX": TokenType.INDEX,
 785        "INET": TokenType.INET,
 786        "INNER": TokenType.INNER,
 787        "INSERT": TokenType.INSERT,
 788        "INTERVAL": TokenType.INTERVAL,
 789        "INTERSECT": TokenType.INTERSECT,
 790        "INTO": TokenType.INTO,
 791        "IS": TokenType.IS,
 792        "ISNULL": TokenType.ISNULL,
 793        "JOIN": TokenType.JOIN,
 794        "KEEP": TokenType.KEEP,
 795        "KILL": TokenType.KILL,
 796        "LATERAL": TokenType.LATERAL,
 797        "LEFT": TokenType.LEFT,
 798        "LIKE": TokenType.LIKE,
 799        "LIMIT": TokenType.LIMIT,
 800        "LOAD": TokenType.LOAD,
 801        "LOCK": TokenType.LOCK,
 802        "MERGE": TokenType.MERGE,
 803        "NAMESPACE": TokenType.NAMESPACE,
 804        "NATURAL": TokenType.NATURAL,
 805        "NEXT": TokenType.NEXT,
 806        "NOT": TokenType.NOT,
 807        "NOTNULL": TokenType.NOTNULL,
 808        "NULL": TokenType.NULL,
 809        "OBJECT": TokenType.OBJECT,
 810        "OFFSET": TokenType.OFFSET,
 811        "ON": TokenType.ON,
 812        "OR": TokenType.OR,
 813        "XOR": TokenType.XOR,
 814        "ORDER BY": TokenType.ORDER_BY,
 815        "ORDINALITY": TokenType.ORDINALITY,
 816        "OUTER": TokenType.OUTER,
 817        "OVER": TokenType.OVER,
 818        "OVERLAPS": TokenType.OVERLAPS,
 819        "OVERWRITE": TokenType.OVERWRITE,
 820        "PARTITION": TokenType.PARTITION,
 821        "PARTITION BY": TokenType.PARTITION_BY,
 822        "PARTITIONED BY": TokenType.PARTITION_BY,
 823        "PARTITIONED_BY": TokenType.PARTITION_BY,
 824        "PERCENT": TokenType.PERCENT,
 825        "PIVOT": TokenType.PIVOT,
 826        "PRAGMA": TokenType.PRAGMA,
 827        "PRIMARY KEY": TokenType.PRIMARY_KEY,
 828        "PROCEDURE": TokenType.PROCEDURE,
 829        "QUALIFY": TokenType.QUALIFY,
 830        "RANGE": TokenType.RANGE,
 831        "RECURSIVE": TokenType.RECURSIVE,
 832        "REGEXP": TokenType.RLIKE,
 833        "RENAME": TokenType.RENAME,
 834        "REPLACE": TokenType.REPLACE,
 835        "RETURNING": TokenType.RETURNING,
 836        "REFERENCES": TokenType.REFERENCES,
 837        "RIGHT": TokenType.RIGHT,
 838        "RLIKE": TokenType.RLIKE,
 839        "ROLLBACK": TokenType.ROLLBACK,
 840        "ROLLUP": TokenType.ROLLUP,
 841        "ROW": TokenType.ROW,
 842        "ROWS": TokenType.ROWS,
 843        "SCHEMA": TokenType.SCHEMA,
 844        "SELECT": TokenType.SELECT,
 845        "SEMI": TokenType.SEMI,
 846        "SESSION": TokenType.SESSION,
 847        "SET": TokenType.SET,
 848        "SETTINGS": TokenType.SETTINGS,
 849        "SHOW": TokenType.SHOW,
 850        "SIMILAR TO": TokenType.SIMILAR_TO,
 851        "SOME": TokenType.SOME,
 852        "SORT BY": TokenType.SORT_BY,
 853        "START WITH": TokenType.START_WITH,
 854        "STRAIGHT_JOIN": TokenType.STRAIGHT_JOIN,
 855        "TABLE": TokenType.TABLE,
 856        "TABLESAMPLE": TokenType.TABLE_SAMPLE,
 857        "TEMP": TokenType.TEMPORARY,
 858        "TEMPORARY": TokenType.TEMPORARY,
 859        "THEN": TokenType.THEN,
 860        "TRUE": TokenType.TRUE,
 861        "TRUNCATE": TokenType.TRUNCATE,
 862        "UNION": TokenType.UNION,
 863        "UNKNOWN": TokenType.UNKNOWN,
 864        "UNNEST": TokenType.UNNEST,
 865        "UNPIVOT": TokenType.UNPIVOT,
 866        "UPDATE": TokenType.UPDATE,
 867        "USE": TokenType.USE,
 868        "USING": TokenType.USING,
 869        "UUID": TokenType.UUID,
 870        "VALUES": TokenType.VALUES,
 871        "VIEW": TokenType.VIEW,
 872        "VOLATILE": TokenType.VOLATILE,
 873        "WHEN": TokenType.WHEN,
 874        "WHERE": TokenType.WHERE,
 875        "WINDOW": TokenType.WINDOW,
 876        "WITH": TokenType.WITH,
 877        "APPLY": TokenType.APPLY,
 878        "ARRAY": TokenType.ARRAY,
 879        "BIT": TokenType.BIT,
 880        "BOOL": TokenType.BOOLEAN,
 881        "BOOLEAN": TokenType.BOOLEAN,
 882        "BYTE": TokenType.TINYINT,
 883        "MEDIUMINT": TokenType.MEDIUMINT,
 884        "INT1": TokenType.TINYINT,
 885        "TINYINT": TokenType.TINYINT,
 886        "INT16": TokenType.SMALLINT,
 887        "SHORT": TokenType.SMALLINT,
 888        "SMALLINT": TokenType.SMALLINT,
 889        "HUGEINT": TokenType.INT128,
 890        "UHUGEINT": TokenType.UINT128,
 891        "INT2": TokenType.SMALLINT,
 892        "INTEGER": TokenType.INT,
 893        "INT": TokenType.INT,
 894        "INT4": TokenType.INT,
 895        "INT32": TokenType.INT,
 896        "INT64": TokenType.BIGINT,
 897        "INT128": TokenType.INT128,
 898        "INT256": TokenType.INT256,
 899        "LONG": TokenType.BIGINT,
 900        "BIGINT": TokenType.BIGINT,
 901        "INT8": TokenType.TINYINT,
 902        "UINT": TokenType.UINT,
 903        "UINT128": TokenType.UINT128,
 904        "UINT256": TokenType.UINT256,
 905        "DEC": TokenType.DECIMAL,
 906        "DECIMAL": TokenType.DECIMAL,
 907        "DECIMAL32": TokenType.DECIMAL32,
 908        "DECIMAL64": TokenType.DECIMAL64,
 909        "DECIMAL128": TokenType.DECIMAL128,
 910        "DECIMAL256": TokenType.DECIMAL256,
 911        "BIGDECIMAL": TokenType.BIGDECIMAL,
 912        "BIGNUMERIC": TokenType.BIGDECIMAL,
 913        "LIST": TokenType.LIST,
 914        "MAP": TokenType.MAP,
 915        "NULLABLE": TokenType.NULLABLE,
 916        "NUMBER": TokenType.DECIMAL,
 917        "NUMERIC": TokenType.DECIMAL,
 918        "FIXED": TokenType.DECIMAL,
 919        "REAL": TokenType.FLOAT,
 920        "FLOAT": TokenType.FLOAT,
 921        "FLOAT4": TokenType.FLOAT,
 922        "FLOAT8": TokenType.DOUBLE,
 923        "DOUBLE": TokenType.DOUBLE,
 924        "DOUBLE PRECISION": TokenType.DOUBLE,
 925        "JSON": TokenType.JSON,
 926        "JSONB": TokenType.JSONB,
 927        "CHAR": TokenType.CHAR,
 928        "CHARACTER": TokenType.CHAR,
 929        "CHAR VARYING": TokenType.VARCHAR,
 930        "CHARACTER VARYING": TokenType.VARCHAR,
 931        "NCHAR": TokenType.NCHAR,
 932        "VARCHAR": TokenType.VARCHAR,
 933        "VARCHAR2": TokenType.VARCHAR,
 934        "NVARCHAR": TokenType.NVARCHAR,
 935        "NVARCHAR2": TokenType.NVARCHAR,
 936        "BPCHAR": TokenType.BPCHAR,
 937        "STR": TokenType.TEXT,
 938        "STRING": TokenType.TEXT,
 939        "TEXT": TokenType.TEXT,
 940        "LONGTEXT": TokenType.LONGTEXT,
 941        "MEDIUMTEXT": TokenType.MEDIUMTEXT,
 942        "TINYTEXT": TokenType.TINYTEXT,
 943        "CLOB": TokenType.TEXT,
 944        "LONGVARCHAR": TokenType.TEXT,
 945        "BINARY": TokenType.BINARY,
 946        "BLOB": TokenType.VARBINARY,
 947        "LONGBLOB": TokenType.LONGBLOB,
 948        "MEDIUMBLOB": TokenType.MEDIUMBLOB,
 949        "TINYBLOB": TokenType.TINYBLOB,
 950        "BYTEA": TokenType.VARBINARY,
 951        "VARBINARY": TokenType.VARBINARY,
 952        "TIME": TokenType.TIME,
 953        "TIMETZ": TokenType.TIMETZ,
 954        "TIMESTAMP": TokenType.TIMESTAMP,
 955        "TIMESTAMPTZ": TokenType.TIMESTAMPTZ,
 956        "TIMESTAMPLTZ": TokenType.TIMESTAMPLTZ,
 957        "TIMESTAMP_LTZ": TokenType.TIMESTAMPLTZ,
 958        "TIMESTAMPNTZ": TokenType.TIMESTAMPNTZ,
 959        "TIMESTAMP_NTZ": TokenType.TIMESTAMPNTZ,
 960        "DATE": TokenType.DATE,
 961        "DATETIME": TokenType.DATETIME,
 962        "INT4RANGE": TokenType.INT4RANGE,
 963        "INT4MULTIRANGE": TokenType.INT4MULTIRANGE,
 964        "INT8RANGE": TokenType.INT8RANGE,
 965        "INT8MULTIRANGE": TokenType.INT8MULTIRANGE,
 966        "NUMRANGE": TokenType.NUMRANGE,
 967        "NUMMULTIRANGE": TokenType.NUMMULTIRANGE,
 968        "TSRANGE": TokenType.TSRANGE,
 969        "TSMULTIRANGE": TokenType.TSMULTIRANGE,
 970        "TSTZRANGE": TokenType.TSTZRANGE,
 971        "TSTZMULTIRANGE": TokenType.TSTZMULTIRANGE,
 972        "DATERANGE": TokenType.DATERANGE,
 973        "DATEMULTIRANGE": TokenType.DATEMULTIRANGE,
 974        "UNIQUE": TokenType.UNIQUE,
 975        "VECTOR": TokenType.VECTOR,
 976        "STRUCT": TokenType.STRUCT,
 977        "SEQUENCE": TokenType.SEQUENCE,
 978        "VARIANT": TokenType.VARIANT,
 979        "ALTER": TokenType.ALTER,
 980        "ANALYZE": TokenType.ANALYZE,
 981        "CALL": TokenType.COMMAND,
 982        "COMMENT": TokenType.COMMENT,
 983        "EXPLAIN": TokenType.COMMAND,
 984        "GRANT": TokenType.GRANT,
 985        "REVOKE": TokenType.REVOKE,
 986        "OPTIMIZE": TokenType.COMMAND,
 987        "PREPARE": TokenType.COMMAND,
 988        "VACUUM": TokenType.COMMAND,
 989        "USER-DEFINED": TokenType.USERDEFINED,
 990        "FOR VERSION": TokenType.VERSION_SNAPSHOT,
 991        "FOR TIMESTAMP": TokenType.TIMESTAMP_SNAPSHOT,
 992    }
 993
 994    WHITE_SPACE: t.Dict[t.Optional[str], TokenType] = {
 995        " ": TokenType.SPACE,
 996        "\t": TokenType.SPACE,
 997        "\n": TokenType.BREAK,
 998        "\r": TokenType.BREAK,
 999    }
1000
1001    COMMANDS = {
1002        TokenType.COMMAND,
1003        TokenType.EXECUTE,
1004        TokenType.FETCH,
1005        TokenType.SHOW,
1006        TokenType.RENAME,
1007    }
1008
1009    COMMAND_PREFIX_TOKENS = {TokenType.SEMICOLON, TokenType.BEGIN}
1010
1011    # Handle numeric literals like in hive (3L = BIGINT)
1012    NUMERIC_LITERALS: t.Dict[str, str] = {}
1013
1014    COMMENTS = ["--", ("/*", "*/")]
1015
1016    __slots__ = (
1017        "sql",
1018        "size",
1019        "tokens",
1020        "dialect",
1021        "use_rs_tokenizer",
1022        "_start",
1023        "_current",
1024        "_line",
1025        "_col",
1026        "_comments",
1027        "_char",
1028        "_end",
1029        "_peek",
1030        "_prev_token_line",
1031        "_rs_dialect_settings",
1032    )
1033
1034    def __init__(
1035        self,
1036        dialect: DialectType = None,
1037        use_rs_tokenizer: t.Optional[bool] = None,
1038        **opts: t.Any,
1039    ) -> None:
1040        from sqlglot.dialects import Dialect
1041
1042        self.dialect = Dialect.get_or_raise(dialect)
1043
1044        # initialize `use_rs_tokenizer`, and allow it to be overwritten per Tokenizer instance
1045        self.use_rs_tokenizer = (
1046            use_rs_tokenizer if use_rs_tokenizer is not None else USE_RS_TOKENIZER
1047        )
1048
1049        if self.use_rs_tokenizer:
1050            self._rs_dialect_settings = RsTokenizerDialectSettings(
1051                unescaped_sequences=self.dialect.UNESCAPED_SEQUENCES,
1052                identifiers_can_start_with_digit=self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT,
1053                numbers_can_be_underscore_separated=self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED,
1054            )
1055
1056        self.reset()
1057
1058    def reset(self) -> None:
1059        self.sql = ""
1060        self.size = 0
1061        self.tokens: t.List[Token] = []
1062        self._start = 0
1063        self._current = 0
1064        self._line = 1
1065        self._col = 0
1066        self._comments: t.List[str] = []
1067
1068        self._char = ""
1069        self._end = False
1070        self._peek = ""
1071        self._prev_token_line = -1
1072
1073    def tokenize(self, sql: str) -> t.List[Token]:
1074        """Returns a list of tokens corresponding to the SQL string `sql`."""
1075        if self.use_rs_tokenizer:
1076            return self.tokenize_rs(sql)
1077
1078        self.reset()
1079        self.sql = sql
1080        self.size = len(sql)
1081
1082        try:
1083            self._scan()
1084        except Exception as e:
1085            start = max(self._current - 50, 0)
1086            end = min(self._current + 50, self.size - 1)
1087            context = self.sql[start:end]
1088            raise TokenError(f"Error tokenizing '{context}'") from e
1089
1090        return self.tokens
1091
1092    def _scan(self, until: t.Optional[t.Callable] = None) -> None:
1093        while self.size and not self._end:
1094            current = self._current
1095
1096            # Skip spaces here rather than iteratively calling advance() for performance reasons
1097            while current < self.size:
1098                char = self.sql[current]
1099
1100                if char.isspace() and (char == " " or char == "\t"):
1101                    current += 1
1102                else:
1103                    break
1104
1105            offset = current - self._current if current > self._current else 1
1106
1107            self._start = current
1108            self._advance(offset)
1109
1110            if not self._char.isspace():
1111                if self._char.isdigit():
1112                    self._scan_number()
1113                elif self._char in self._IDENTIFIERS:
1114                    self._scan_identifier(self._IDENTIFIERS[self._char])
1115                else:
1116                    self._scan_keywords()
1117
1118            if until and until():
1119                break
1120
1121        if self.tokens and self._comments:
1122            self.tokens[-1].comments.extend(self._comments)
1123
1124    def _chars(self, size: int) -> str:
1125        if size == 1:
1126            return self._char
1127
1128        start = self._current - 1
1129        end = start + size
1130
1131        return self.sql[start:end] if end <= self.size else ""
1132
1133    def _advance(self, i: int = 1, alnum: bool = False) -> None:
1134        if self.WHITE_SPACE.get(self._char) is TokenType.BREAK:
1135            # Ensures we don't count an extra line if we get a \r\n line break sequence
1136            if not (self._char == "\r" and self._peek == "\n"):
1137                self._col = i
1138                self._line += 1
1139        else:
1140            self._col += i
1141
1142        self._current += i
1143        self._end = self._current >= self.size
1144        self._char = self.sql[self._current - 1]
1145        self._peek = "" if self._end else self.sql[self._current]
1146
1147        if alnum and self._char.isalnum():
1148            # Here we use local variables instead of attributes for better performance
1149            _col = self._col
1150            _current = self._current
1151            _end = self._end
1152            _peek = self._peek
1153
1154            while _peek.isalnum():
1155                _col += 1
1156                _current += 1
1157                _end = _current >= self.size
1158                _peek = "" if _end else self.sql[_current]
1159
1160            self._col = _col
1161            self._current = _current
1162            self._end = _end
1163            self._peek = _peek
1164            self._char = self.sql[_current - 1]
1165
1166    @property
1167    def _text(self) -> str:
1168        return self.sql[self._start : self._current]
1169
1170    def _add(self, token_type: TokenType, text: t.Optional[str] = None) -> None:
1171        self._prev_token_line = self._line
1172
1173        if self._comments and token_type == TokenType.SEMICOLON and self.tokens:
1174            self.tokens[-1].comments.extend(self._comments)
1175            self._comments = []
1176
1177        self.tokens.append(
1178            Token(
1179                token_type,
1180                text=self._text if text is None else text,
1181                line=self._line,
1182                col=self._col,
1183                start=self._start,
1184                end=self._current - 1,
1185                comments=self._comments,
1186            )
1187        )
1188        self._comments = []
1189
1190        # If we have either a semicolon or a begin token before the command's token, we'll parse
1191        # whatever follows the command's token as a string
1192        if (
1193            token_type in self.COMMANDS
1194            and self._peek != ";"
1195            and (len(self.tokens) == 1 or self.tokens[-2].token_type in self.COMMAND_PREFIX_TOKENS)
1196        ):
1197            start = self._current
1198            tokens = len(self.tokens)
1199            self._scan(lambda: self._peek == ";")
1200            self.tokens = self.tokens[:tokens]
1201            text = self.sql[start : self._current].strip()
1202            if text:
1203                self._add(TokenType.STRING, text)
1204
1205    def _scan_keywords(self) -> None:
1206        size = 0
1207        word = None
1208        chars = self._text
1209        char = chars
1210        prev_space = False
1211        skip = False
1212        trie = self._KEYWORD_TRIE
1213        single_token = char in self.SINGLE_TOKENS
1214
1215        while chars:
1216            if skip:
1217                result = TrieResult.PREFIX
1218            else:
1219                result, trie = in_trie(trie, char.upper())
1220
1221            if result == TrieResult.FAILED:
1222                break
1223            if result == TrieResult.EXISTS:
1224                word = chars
1225
1226            end = self._current + size
1227            size += 1
1228
1229            if end < self.size:
1230                char = self.sql[end]
1231                single_token = single_token or char in self.SINGLE_TOKENS
1232                is_space = char.isspace()
1233
1234                if not is_space or not prev_space:
1235                    if is_space:
1236                        char = " "
1237                    chars += char
1238                    prev_space = is_space
1239                    skip = False
1240                else:
1241                    skip = True
1242            else:
1243                char = ""
1244                break
1245
1246        if word:
1247            if self._scan_string(word):
1248                return
1249            if self._scan_comment(word):
1250                return
1251            if prev_space or single_token or not char:
1252                self._advance(size - 1)
1253                word = word.upper()
1254                self._add(self.KEYWORDS[word], text=word)
1255                return
1256
1257        if self._char in self.SINGLE_TOKENS:
1258            self._add(self.SINGLE_TOKENS[self._char], text=self._char)
1259            return
1260
1261        self._scan_var()
1262
1263    def _scan_comment(self, comment_start: str) -> bool:
1264        if comment_start not in self._COMMENTS:
1265            return False
1266
1267        comment_start_line = self._line
1268        comment_start_size = len(comment_start)
1269        comment_end = self._COMMENTS[comment_start]
1270
1271        if comment_end:
1272            # Skip the comment's start delimiter
1273            self._advance(comment_start_size)
1274
1275            comment_count = 1
1276            comment_end_size = len(comment_end)
1277
1278            while not self._end:
1279                if self._chars(comment_end_size) == comment_end:
1280                    comment_count -= 1
1281                    if not comment_count:
1282                        break
1283
1284                self._advance(alnum=True)
1285
1286                # Nested comments are allowed by some dialects, e.g. databricks, duckdb, postgres
1287                if (
1288                    self.NESTED_COMMENTS
1289                    and not self._end
1290                    and self._chars(comment_end_size) == comment_start
1291                ):
1292                    self._advance(comment_start_size)
1293                    comment_count += 1
1294
1295            self._comments.append(self._text[comment_start_size : -comment_end_size + 1])
1296            self._advance(comment_end_size - 1)
1297        else:
1298            while not self._end and self.WHITE_SPACE.get(self._peek) is not TokenType.BREAK:
1299                self._advance(alnum=True)
1300            self._comments.append(self._text[comment_start_size:])
1301
1302        if (
1303            comment_start == self.HINT_START
1304            and self.tokens
1305            and self.tokens[-1].token_type in self.TOKENS_PRECEDING_HINT
1306        ):
1307            self._add(TokenType.HINT)
1308
1309        # Leading comment is attached to the succeeding token, whilst trailing comment to the preceding.
1310        # Multiple consecutive comments are preserved by appending them to the current comments list.
1311        if comment_start_line == self._prev_token_line:
1312            self.tokens[-1].comments.extend(self._comments)
1313            self._comments = []
1314            self._prev_token_line = self._line
1315
1316        return True
1317
1318    def _scan_number(self) -> None:
1319        if self._char == "0":
1320            peek = self._peek.upper()
1321            if peek == "B":
1322                return self._scan_bits() if self.BIT_STRINGS else self._add(TokenType.NUMBER)
1323            elif peek == "X":
1324                return self._scan_hex() if self.HEX_STRINGS else self._add(TokenType.NUMBER)
1325
1326        decimal = False
1327        scientific = 0
1328
1329        while True:
1330            if self._peek.isdigit():
1331                self._advance()
1332            elif self._peek == "." and not decimal:
1333                if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER:
1334                    return self._add(TokenType.NUMBER)
1335                decimal = True
1336                self._advance()
1337            elif self._peek in ("-", "+") and scientific == 1:
1338                scientific += 1
1339                self._advance()
1340            elif self._peek.upper() == "E" and not scientific:
1341                scientific += 1
1342                self._advance()
1343            elif self._peek.isidentifier():
1344                number_text = self._text
1345                literal = ""
1346
1347                while self._peek.strip() and self._peek not in self.SINGLE_TOKENS:
1348                    literal += self._peek
1349                    self._advance()
1350
1351                token_type = self.KEYWORDS.get(self.NUMERIC_LITERALS.get(literal.upper(), ""))
1352
1353                if token_type:
1354                    self._add(TokenType.NUMBER, number_text)
1355                    self._add(TokenType.DCOLON, "::")
1356                    return self._add(token_type, literal)
1357                else:
1358                    replaced = literal.replace("_", "")
1359                    if self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED and replaced.isdigit():
1360                        return self._add(TokenType.NUMBER, number_text + replaced)
1361                    if self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT:
1362                        return self._add(TokenType.VAR)
1363
1364                self._advance(-len(literal))
1365                return self._add(TokenType.NUMBER, number_text)
1366            else:
1367                return self._add(TokenType.NUMBER)
1368
1369    def _scan_bits(self) -> None:
1370        self._advance()
1371        value = self._extract_value()
1372        try:
1373            # If `value` can't be converted to a binary, fallback to tokenizing it as an identifier
1374            int(value, 2)
1375            self._add(TokenType.BIT_STRING, value[2:])  # Drop the 0b
1376        except ValueError:
1377            self._add(TokenType.IDENTIFIER)
1378
1379    def _scan_hex(self) -> None:
1380        self._advance()
1381        value = self._extract_value()
1382        try:
1383            # If `value` can't be converted to a hex, fallback to tokenizing it as an identifier
1384            int(value, 16)
1385            self._add(TokenType.HEX_STRING, value[2:])  # Drop the 0x
1386        except ValueError:
1387            self._add(TokenType.IDENTIFIER)
1388
1389    def _extract_value(self) -> str:
1390        while True:
1391            char = self._peek.strip()
1392            if char and char not in self.SINGLE_TOKENS:
1393                self._advance(alnum=True)
1394            else:
1395                break
1396
1397        return self._text
1398
1399    def _scan_string(self, start: str) -> bool:
1400        base = None
1401        token_type = TokenType.STRING
1402
1403        if start in self._QUOTES:
1404            end = self._QUOTES[start]
1405        elif start in self._FORMAT_STRINGS:
1406            end, token_type = self._FORMAT_STRINGS[start]
1407
1408            if token_type == TokenType.HEX_STRING:
1409                base = 16
1410            elif token_type == TokenType.BIT_STRING:
1411                base = 2
1412            elif token_type == TokenType.HEREDOC_STRING:
1413                self._advance()
1414
1415                if self._char == end:
1416                    tag = ""
1417                else:
1418                    tag = self._extract_string(
1419                        end,
1420                        raw_string=True,
1421                        raise_unmatched=not self.HEREDOC_TAG_IS_IDENTIFIER,
1422                    )
1423
1424                if tag and self.HEREDOC_TAG_IS_IDENTIFIER and (self._end or not tag.isidentifier()):
1425                    if not self._end:
1426                        self._advance(-1)
1427
1428                    self._advance(-len(tag))
1429                    self._add(self.HEREDOC_STRING_ALTERNATIVE)
1430                    return True
1431
1432                end = f"{start}{tag}{end}"
1433        else:
1434            return False
1435
1436        self._advance(len(start))
1437        text = self._extract_string(end, raw_string=token_type == TokenType.RAW_STRING)
1438
1439        if base and text:
1440            try:
1441                int(text, base)
1442            except Exception:
1443                raise TokenError(
1444                    f"Numeric string contains invalid characters from {self._line}:{self._start}"
1445                )
1446
1447        self._add(token_type, text)
1448        return True
1449
1450    def _scan_identifier(self, identifier_end: str) -> None:
1451        self._advance()
1452        text = self._extract_string(
1453            identifier_end, escapes=self._IDENTIFIER_ESCAPES | {identifier_end}
1454        )
1455        self._add(TokenType.IDENTIFIER, text)
1456
1457    def _scan_var(self) -> None:
1458        while True:
1459            char = self._peek.strip()
1460            if char and (char in self.VAR_SINGLE_TOKENS or char not in self.SINGLE_TOKENS):
1461                self._advance(alnum=True)
1462            else:
1463                break
1464
1465        self._add(
1466            TokenType.VAR
1467            if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER
1468            else self.KEYWORDS.get(self._text.upper(), TokenType.VAR)
1469        )
1470
1471    def _extract_string(
1472        self,
1473        delimiter: str,
1474        escapes: t.Optional[t.Set[str]] = None,
1475        raw_string: bool = False,
1476        raise_unmatched: bool = True,
1477    ) -> str:
1478        text = ""
1479        delim_size = len(delimiter)
1480        escapes = self._STRING_ESCAPES if escapes is None else escapes
1481
1482        while True:
1483            if (
1484                not raw_string
1485                and self.dialect.UNESCAPED_SEQUENCES
1486                and self._peek
1487                and self._char in self.STRING_ESCAPES
1488            ):
1489                unescaped_sequence = self.dialect.UNESCAPED_SEQUENCES.get(self._char + self._peek)
1490                if unescaped_sequence:
1491                    self._advance(2)
1492                    text += unescaped_sequence
1493                    continue
1494            if (
1495                (self.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS or not raw_string)
1496                and self._char in escapes
1497                and (self._peek == delimiter or self._peek in escapes)
1498                and (self._char not in self._QUOTES or self._char == self._peek)
1499            ):
1500                if self._peek == delimiter:
1501                    text += self._peek
1502                else:
1503                    text += self._char + self._peek
1504
1505                if self._current + 1 < self.size:
1506                    self._advance(2)
1507                else:
1508                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._current}")
1509            else:
1510                if self._chars(delim_size) == delimiter:
1511                    if delim_size > 1:
1512                        self._advance(delim_size - 1)
1513                    break
1514
1515                if self._end:
1516                    if not raise_unmatched:
1517                        return text + self._char
1518
1519                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._start}")
1520
1521                current = self._current - 1
1522                self._advance(alnum=True)
1523                text += self.sql[current : self._current - 1]
1524
1525        return text
1526
1527    def tokenize_rs(self, sql: str) -> t.List[Token]:
1528        if not self._RS_TOKENIZER:
1529            raise SqlglotError("Rust tokenizer is not available")
1530
1531        tokens, error_msg = self._RS_TOKENIZER.tokenize(sql, self._rs_dialect_settings)
1532        for token in tokens:
1533            token.token_type = _ALL_TOKEN_TYPES[token.token_type_index]
1534
1535        # Setting this here so partial token lists can be inspected even if there is a failure
1536        self.tokens = tokens
1537
1538        if error_msg is not None:
1539            raise TokenError(error_msg)
1540
1541        return tokens