sqlglot.tokens

   1from __future__ import annotations
   2
   3import os
   4import typing as t
   5from enum import auto
   6
   7from sqlglot.errors import SqlglotError, TokenError
   8from sqlglot.helper import AutoName
   9from sqlglot.trie import TrieResult, in_trie, new_trie
  10
  11if t.TYPE_CHECKING:
  12    from sqlglot.dialects.dialect import DialectType
  13
  14
  15try:
  16    from sqlglotrs import (  # type: ignore
  17        Tokenizer as RsTokenizer,
  18        TokenizerDialectSettings as RsTokenizerDialectSettings,
  19        TokenizerSettings as RsTokenizerSettings,
  20        TokenTypeSettings as RsTokenTypeSettings,
  21    )
  22
  23    USE_RS_TOKENIZER = os.environ.get("SQLGLOTRS_TOKENIZER", "1") == "1"
  24except ImportError:
  25    USE_RS_TOKENIZER = False
  26
  27
  28class TokenType(AutoName):
  29    L_PAREN = auto()
  30    R_PAREN = auto()
  31    L_BRACKET = auto()
  32    R_BRACKET = auto()
  33    L_BRACE = auto()
  34    R_BRACE = auto()
  35    COMMA = auto()
  36    DOT = auto()
  37    DASH = auto()
  38    PLUS = auto()
  39    COLON = auto()
  40    DCOLON = auto()
  41    DQMARK = auto()
  42    SEMICOLON = auto()
  43    STAR = auto()
  44    BACKSLASH = auto()
  45    SLASH = auto()
  46    LT = auto()
  47    LTE = auto()
  48    GT = auto()
  49    GTE = auto()
  50    NOT = auto()
  51    EQ = auto()
  52    NEQ = auto()
  53    NULLSAFE_EQ = auto()
  54    COLON_EQ = auto()
  55    AND = auto()
  56    OR = auto()
  57    AMP = auto()
  58    DPIPE = auto()
  59    PIPE = auto()
  60    PIPE_SLASH = auto()
  61    DPIPE_SLASH = auto()
  62    CARET = auto()
  63    CARET_AT = auto()
  64    TILDA = auto()
  65    ARROW = auto()
  66    DARROW = auto()
  67    FARROW = auto()
  68    HASH = auto()
  69    HASH_ARROW = auto()
  70    DHASH_ARROW = auto()
  71    LR_ARROW = auto()
  72    DAT = auto()
  73    LT_AT = auto()
  74    AT_GT = auto()
  75    DOLLAR = auto()
  76    PARAMETER = auto()
  77    SESSION_PARAMETER = auto()
  78    DAMP = auto()
  79    XOR = auto()
  80    DSTAR = auto()
  81
  82    BLOCK_START = auto()
  83    BLOCK_END = auto()
  84
  85    SPACE = auto()
  86    BREAK = auto()
  87
  88    STRING = auto()
  89    NUMBER = auto()
  90    IDENTIFIER = auto()
  91    DATABASE = auto()
  92    COLUMN = auto()
  93    COLUMN_DEF = auto()
  94    SCHEMA = auto()
  95    TABLE = auto()
  96    WAREHOUSE = auto()
  97    STREAMLIT = auto()
  98    VAR = auto()
  99    BIT_STRING = auto()
 100    HEX_STRING = auto()
 101    BYTE_STRING = auto()
 102    NATIONAL_STRING = auto()
 103    RAW_STRING = auto()
 104    HEREDOC_STRING = auto()
 105    UNICODE_STRING = auto()
 106
 107    # types
 108    BIT = auto()
 109    BOOLEAN = auto()
 110    TINYINT = auto()
 111    UTINYINT = auto()
 112    SMALLINT = auto()
 113    USMALLINT = auto()
 114    MEDIUMINT = auto()
 115    UMEDIUMINT = auto()
 116    INT = auto()
 117    UINT = auto()
 118    BIGINT = auto()
 119    UBIGINT = auto()
 120    INT128 = auto()
 121    UINT128 = auto()
 122    INT256 = auto()
 123    UINT256 = auto()
 124    FLOAT = auto()
 125    DOUBLE = auto()
 126    DECIMAL = auto()
 127    DECIMAL32 = auto()
 128    DECIMAL64 = auto()
 129    DECIMAL128 = auto()
 130    UDECIMAL = auto()
 131    BIGDECIMAL = auto()
 132    CHAR = auto()
 133    NCHAR = auto()
 134    VARCHAR = auto()
 135    NVARCHAR = auto()
 136    BPCHAR = auto()
 137    TEXT = auto()
 138    MEDIUMTEXT = auto()
 139    LONGTEXT = auto()
 140    MEDIUMBLOB = auto()
 141    LONGBLOB = auto()
 142    TINYBLOB = auto()
 143    TINYTEXT = auto()
 144    NAME = auto()
 145    BINARY = auto()
 146    VARBINARY = auto()
 147    JSON = auto()
 148    JSONB = auto()
 149    TIME = auto()
 150    TIMETZ = auto()
 151    TIMESTAMP = auto()
 152    TIMESTAMPTZ = auto()
 153    TIMESTAMPLTZ = auto()
 154    TIMESTAMPNTZ = auto()
 155    TIMESTAMP_S = auto()
 156    TIMESTAMP_MS = auto()
 157    TIMESTAMP_NS = auto()
 158    DATETIME = auto()
 159    DATETIME64 = auto()
 160    DATE = auto()
 161    DATE32 = auto()
 162    INT4RANGE = auto()
 163    INT4MULTIRANGE = auto()
 164    INT8RANGE = auto()
 165    INT8MULTIRANGE = auto()
 166    NUMRANGE = auto()
 167    NUMMULTIRANGE = auto()
 168    TSRANGE = auto()
 169    TSMULTIRANGE = auto()
 170    TSTZRANGE = auto()
 171    TSTZMULTIRANGE = auto()
 172    DATERANGE = auto()
 173    DATEMULTIRANGE = auto()
 174    UUID = auto()
 175    GEOGRAPHY = auto()
 176    NULLABLE = auto()
 177    GEOMETRY = auto()
 178    HLLSKETCH = auto()
 179    HSTORE = auto()
 180    SUPER = auto()
 181    SERIAL = auto()
 182    SMALLSERIAL = auto()
 183    BIGSERIAL = auto()
 184    XML = auto()
 185    YEAR = auto()
 186    UNIQUEIDENTIFIER = auto()
 187    USERDEFINED = auto()
 188    MONEY = auto()
 189    SMALLMONEY = auto()
 190    ROWVERSION = auto()
 191    IMAGE = auto()
 192    VARIANT = auto()
 193    OBJECT = auto()
 194    INET = auto()
 195    IPADDRESS = auto()
 196    IPPREFIX = auto()
 197    IPV4 = auto()
 198    IPV6 = auto()
 199    ENUM = auto()
 200    ENUM8 = auto()
 201    ENUM16 = auto()
 202    FIXEDSTRING = auto()
 203    LOWCARDINALITY = auto()
 204    NESTED = auto()
 205    AGGREGATEFUNCTION = auto()
 206    SIMPLEAGGREGATEFUNCTION = auto()
 207    TDIGEST = auto()
 208    UNKNOWN = auto()
 209    VECTOR = auto()
 210
 211    # keywords
 212    ALIAS = auto()
 213    ALTER = auto()
 214    ALWAYS = auto()
 215    ALL = auto()
 216    ANTI = auto()
 217    ANY = auto()
 218    APPLY = auto()
 219    ARRAY = auto()
 220    ASC = auto()
 221    ASOF = auto()
 222    AUTO_INCREMENT = auto()
 223    BEGIN = auto()
 224    BETWEEN = auto()
 225    BULK_COLLECT_INTO = auto()
 226    CACHE = auto()
 227    CASE = auto()
 228    CHARACTER_SET = auto()
 229    CLUSTER_BY = auto()
 230    COLLATE = auto()
 231    COMMAND = auto()
 232    COMMENT = auto()
 233    COMMIT = auto()
 234    CONNECT_BY = auto()
 235    CONSTRAINT = auto()
 236    COPY = auto()
 237    CREATE = auto()
 238    CROSS = auto()
 239    CUBE = auto()
 240    CURRENT_DATE = auto()
 241    CURRENT_DATETIME = auto()
 242    CURRENT_TIME = auto()
 243    CURRENT_TIMESTAMP = auto()
 244    CURRENT_USER = auto()
 245    DECLARE = auto()
 246    DEFAULT = auto()
 247    DELETE = auto()
 248    DESC = auto()
 249    DESCRIBE = auto()
 250    DICTIONARY = auto()
 251    DISTINCT = auto()
 252    DISTRIBUTE_BY = auto()
 253    DIV = auto()
 254    DROP = auto()
 255    ELSE = auto()
 256    END = auto()
 257    ESCAPE = auto()
 258    EXCEPT = auto()
 259    EXECUTE = auto()
 260    EXISTS = auto()
 261    FALSE = auto()
 262    FETCH = auto()
 263    FILTER = auto()
 264    FINAL = auto()
 265    FIRST = auto()
 266    FOR = auto()
 267    FORCE = auto()
 268    FOREIGN_KEY = auto()
 269    FORMAT = auto()
 270    FROM = auto()
 271    FULL = auto()
 272    FUNCTION = auto()
 273    GLOB = auto()
 274    GLOBAL = auto()
 275    GRANT = auto()
 276    GROUP_BY = auto()
 277    GROUPING_SETS = auto()
 278    HAVING = auto()
 279    HINT = auto()
 280    IGNORE = auto()
 281    ILIKE = auto()
 282    ILIKE_ANY = auto()
 283    IN = auto()
 284    INDEX = auto()
 285    INNER = auto()
 286    INSERT = auto()
 287    INTERSECT = auto()
 288    INTERVAL = auto()
 289    INTO = auto()
 290    INTRODUCER = auto()
 291    IRLIKE = auto()
 292    IS = auto()
 293    ISNULL = auto()
 294    JOIN = auto()
 295    JOIN_MARKER = auto()
 296    KEEP = auto()
 297    KEY = auto()
 298    KILL = auto()
 299    LANGUAGE = auto()
 300    LATERAL = auto()
 301    LEFT = auto()
 302    LIKE = auto()
 303    LIKE_ANY = auto()
 304    LIMIT = auto()
 305    LIST = auto()
 306    LOAD = auto()
 307    LOCK = auto()
 308    MAP = auto()
 309    MATCH_CONDITION = auto()
 310    MATCH_RECOGNIZE = auto()
 311    MEMBER_OF = auto()
 312    MERGE = auto()
 313    MOD = auto()
 314    MODEL = auto()
 315    NATURAL = auto()
 316    NEXT = auto()
 317    NOTNULL = auto()
 318    NULL = auto()
 319    OBJECT_IDENTIFIER = auto()
 320    OFFSET = auto()
 321    ON = auto()
 322    ONLY = auto()
 323    OPERATOR = auto()
 324    ORDER_BY = auto()
 325    ORDER_SIBLINGS_BY = auto()
 326    ORDERED = auto()
 327    ORDINALITY = auto()
 328    OUTER = auto()
 329    OVER = auto()
 330    OVERLAPS = auto()
 331    OVERWRITE = auto()
 332    PARTITION = auto()
 333    PARTITION_BY = auto()
 334    PERCENT = auto()
 335    PIVOT = auto()
 336    PLACEHOLDER = auto()
 337    POSITIONAL = auto()
 338    PRAGMA = auto()
 339    PREWHERE = auto()
 340    PRIMARY_KEY = auto()
 341    PROCEDURE = auto()
 342    PROPERTIES = auto()
 343    PSEUDO_TYPE = auto()
 344    QUALIFY = auto()
 345    QUOTE = auto()
 346    RANGE = auto()
 347    RECURSIVE = auto()
 348    REFRESH = auto()
 349    RENAME = auto()
 350    REPLACE = auto()
 351    RETURNING = auto()
 352    REFERENCES = auto()
 353    RIGHT = auto()
 354    RLIKE = auto()
 355    ROLLBACK = auto()
 356    ROLLUP = auto()
 357    ROW = auto()
 358    ROWS = auto()
 359    SELECT = auto()
 360    SEMI = auto()
 361    SEPARATOR = auto()
 362    SEQUENCE = auto()
 363    SERDE_PROPERTIES = auto()
 364    SET = auto()
 365    SETTINGS = auto()
 366    SHOW = auto()
 367    SIMILAR_TO = auto()
 368    SOME = auto()
 369    SORT_BY = auto()
 370    START_WITH = auto()
 371    STORAGE_INTEGRATION = auto()
 372    STRAIGHT_JOIN = auto()
 373    STRUCT = auto()
 374    SUMMARIZE = auto()
 375    TABLE_SAMPLE = auto()
 376    TAG = auto()
 377    TEMPORARY = auto()
 378    TOP = auto()
 379    THEN = auto()
 380    TRUE = auto()
 381    TRUNCATE = auto()
 382    UNCACHE = auto()
 383    UNION = auto()
 384    UNNEST = auto()
 385    UNPIVOT = auto()
 386    UPDATE = auto()
 387    USE = auto()
 388    USING = auto()
 389    VALUES = auto()
 390    VIEW = auto()
 391    VOLATILE = auto()
 392    WHEN = auto()
 393    WHERE = auto()
 394    WINDOW = auto()
 395    WITH = auto()
 396    UNIQUE = auto()
 397    VERSION_SNAPSHOT = auto()
 398    TIMESTAMP_SNAPSHOT = auto()
 399    OPTION = auto()
 400
 401
 402_ALL_TOKEN_TYPES = list(TokenType)
 403_TOKEN_TYPE_TO_INDEX = {token_type: i for i, token_type in enumerate(_ALL_TOKEN_TYPES)}
 404
 405
 406class Token:
 407    __slots__ = ("token_type", "text", "line", "col", "start", "end", "comments")
 408
 409    @classmethod
 410    def number(cls, number: int) -> Token:
 411        """Returns a NUMBER token with `number` as its text."""
 412        return cls(TokenType.NUMBER, str(number))
 413
 414    @classmethod
 415    def string(cls, string: str) -> Token:
 416        """Returns a STRING token with `string` as its text."""
 417        return cls(TokenType.STRING, string)
 418
 419    @classmethod
 420    def identifier(cls, identifier: str) -> Token:
 421        """Returns an IDENTIFIER token with `identifier` as its text."""
 422        return cls(TokenType.IDENTIFIER, identifier)
 423
 424    @classmethod
 425    def var(cls, var: str) -> Token:
 426        """Returns an VAR token with `var` as its text."""
 427        return cls(TokenType.VAR, var)
 428
 429    def __init__(
 430        self,
 431        token_type: TokenType,
 432        text: str,
 433        line: int = 1,
 434        col: int = 1,
 435        start: int = 0,
 436        end: int = 0,
 437        comments: t.Optional[t.List[str]] = None,
 438    ) -> None:
 439        """Token initializer.
 440
 441        Args:
 442            token_type: The TokenType Enum.
 443            text: The text of the token.
 444            line: The line that the token ends on.
 445            col: The column that the token ends on.
 446            start: The start index of the token.
 447            end: The ending index of the token.
 448            comments: The comments to attach to the token.
 449        """
 450        self.token_type = token_type
 451        self.text = text
 452        self.line = line
 453        self.col = col
 454        self.start = start
 455        self.end = end
 456        self.comments = [] if comments is None else comments
 457
 458    def __repr__(self) -> str:
 459        attributes = ", ".join(f"{k}: {getattr(self, k)}" for k in self.__slots__)
 460        return f"<Token {attributes}>"
 461
 462
 463class _Tokenizer(type):
 464    def __new__(cls, clsname, bases, attrs):
 465        klass = super().__new__(cls, clsname, bases, attrs)
 466
 467        def _convert_quotes(arr: t.List[str | t.Tuple[str, str]]) -> t.Dict[str, str]:
 468            return dict(
 469                (item, item) if isinstance(item, str) else (item[0], item[1]) for item in arr
 470            )
 471
 472        def _quotes_to_format(
 473            token_type: TokenType, arr: t.List[str | t.Tuple[str, str]]
 474        ) -> t.Dict[str, t.Tuple[str, TokenType]]:
 475            return {k: (v, token_type) for k, v in _convert_quotes(arr).items()}
 476
 477        klass._QUOTES = _convert_quotes(klass.QUOTES)
 478        klass._IDENTIFIERS = _convert_quotes(klass.IDENTIFIERS)
 479
 480        klass._FORMAT_STRINGS = {
 481            **{
 482                p + s: (e, TokenType.NATIONAL_STRING)
 483                for s, e in klass._QUOTES.items()
 484                for p in ("n", "N")
 485            },
 486            **_quotes_to_format(TokenType.BIT_STRING, klass.BIT_STRINGS),
 487            **_quotes_to_format(TokenType.BYTE_STRING, klass.BYTE_STRINGS),
 488            **_quotes_to_format(TokenType.HEX_STRING, klass.HEX_STRINGS),
 489            **_quotes_to_format(TokenType.RAW_STRING, klass.RAW_STRINGS),
 490            **_quotes_to_format(TokenType.HEREDOC_STRING, klass.HEREDOC_STRINGS),
 491            **_quotes_to_format(TokenType.UNICODE_STRING, klass.UNICODE_STRINGS),
 492        }
 493
 494        klass._STRING_ESCAPES = set(klass.STRING_ESCAPES)
 495        klass._IDENTIFIER_ESCAPES = set(klass.IDENTIFIER_ESCAPES)
 496        klass._COMMENTS = {
 497            **dict(
 498                (comment, None) if isinstance(comment, str) else (comment[0], comment[1])
 499                for comment in klass.COMMENTS
 500            ),
 501            "{#": "#}",  # Ensure Jinja comments are tokenized correctly in all dialects
 502        }
 503
 504        klass._KEYWORD_TRIE = new_trie(
 505            key.upper()
 506            for key in (
 507                *klass.KEYWORDS,
 508                *klass._COMMENTS,
 509                *klass._QUOTES,
 510                *klass._FORMAT_STRINGS,
 511            )
 512            if " " in key or any(single in key for single in klass.SINGLE_TOKENS)
 513        )
 514
 515        if USE_RS_TOKENIZER:
 516            settings = RsTokenizerSettings(
 517                white_space={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.WHITE_SPACE.items()},
 518                single_tokens={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.SINGLE_TOKENS.items()},
 519                keywords={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.KEYWORDS.items()},
 520                numeric_literals=klass.NUMERIC_LITERALS,
 521                identifiers=klass._IDENTIFIERS,
 522                identifier_escapes=klass._IDENTIFIER_ESCAPES,
 523                string_escapes=klass._STRING_ESCAPES,
 524                quotes=klass._QUOTES,
 525                format_strings={
 526                    k: (v1, _TOKEN_TYPE_TO_INDEX[v2])
 527                    for k, (v1, v2) in klass._FORMAT_STRINGS.items()
 528                },
 529                has_bit_strings=bool(klass.BIT_STRINGS),
 530                has_hex_strings=bool(klass.HEX_STRINGS),
 531                comments=klass._COMMENTS,
 532                var_single_tokens=klass.VAR_SINGLE_TOKENS,
 533                commands={_TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMANDS},
 534                command_prefix_tokens={
 535                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMAND_PREFIX_TOKENS
 536                },
 537                heredoc_tag_is_identifier=klass.HEREDOC_TAG_IS_IDENTIFIER,
 538                string_escapes_allowed_in_raw_strings=klass.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS,
 539                nested_comments=klass.NESTED_COMMENTS,
 540            )
 541            token_types = RsTokenTypeSettings(
 542                bit_string=_TOKEN_TYPE_TO_INDEX[TokenType.BIT_STRING],
 543                break_=_TOKEN_TYPE_TO_INDEX[TokenType.BREAK],
 544                dcolon=_TOKEN_TYPE_TO_INDEX[TokenType.DCOLON],
 545                heredoc_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEREDOC_STRING],
 546                raw_string=_TOKEN_TYPE_TO_INDEX[TokenType.RAW_STRING],
 547                hex_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEX_STRING],
 548                identifier=_TOKEN_TYPE_TO_INDEX[TokenType.IDENTIFIER],
 549                number=_TOKEN_TYPE_TO_INDEX[TokenType.NUMBER],
 550                parameter=_TOKEN_TYPE_TO_INDEX[TokenType.PARAMETER],
 551                semicolon=_TOKEN_TYPE_TO_INDEX[TokenType.SEMICOLON],
 552                string=_TOKEN_TYPE_TO_INDEX[TokenType.STRING],
 553                var=_TOKEN_TYPE_TO_INDEX[TokenType.VAR],
 554                heredoc_string_alternative=_TOKEN_TYPE_TO_INDEX[klass.HEREDOC_STRING_ALTERNATIVE],
 555            )
 556            klass._RS_TOKENIZER = RsTokenizer(settings, token_types)
 557        else:
 558            klass._RS_TOKENIZER = None
 559
 560        return klass
 561
 562
 563class Tokenizer(metaclass=_Tokenizer):
 564    SINGLE_TOKENS = {
 565        "(": TokenType.L_PAREN,
 566        ")": TokenType.R_PAREN,
 567        "[": TokenType.L_BRACKET,
 568        "]": TokenType.R_BRACKET,
 569        "{": TokenType.L_BRACE,
 570        "}": TokenType.R_BRACE,
 571        "&": TokenType.AMP,
 572        "^": TokenType.CARET,
 573        ":": TokenType.COLON,
 574        ",": TokenType.COMMA,
 575        ".": TokenType.DOT,
 576        "-": TokenType.DASH,
 577        "=": TokenType.EQ,
 578        ">": TokenType.GT,
 579        "<": TokenType.LT,
 580        "%": TokenType.MOD,
 581        "!": TokenType.NOT,
 582        "|": TokenType.PIPE,
 583        "+": TokenType.PLUS,
 584        ";": TokenType.SEMICOLON,
 585        "/": TokenType.SLASH,
 586        "\\": TokenType.BACKSLASH,
 587        "*": TokenType.STAR,
 588        "~": TokenType.TILDA,
 589        "?": TokenType.PLACEHOLDER,
 590        "@": TokenType.PARAMETER,
 591        "#": TokenType.HASH,
 592        # Used for breaking a var like x'y' but nothing else the token type doesn't matter
 593        "'": TokenType.UNKNOWN,
 594        "`": TokenType.UNKNOWN,
 595        '"': TokenType.UNKNOWN,
 596    }
 597
 598    BIT_STRINGS: t.List[str | t.Tuple[str, str]] = []
 599    BYTE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 600    HEX_STRINGS: t.List[str | t.Tuple[str, str]] = []
 601    RAW_STRINGS: t.List[str | t.Tuple[str, str]] = []
 602    HEREDOC_STRINGS: t.List[str | t.Tuple[str, str]] = []
 603    UNICODE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 604    IDENTIFIERS: t.List[str | t.Tuple[str, str]] = ['"']
 605    IDENTIFIER_ESCAPES = ['"']
 606    QUOTES: t.List[t.Tuple[str, str] | str] = ["'"]
 607    STRING_ESCAPES = ["'"]
 608    VAR_SINGLE_TOKENS: t.Set[str] = set()
 609
 610    # Whether the heredoc tags follow the same lexical rules as unquoted identifiers
 611    HEREDOC_TAG_IS_IDENTIFIER = False
 612
 613    # Token that we'll generate as a fallback if the heredoc prefix doesn't correspond to a heredoc
 614    HEREDOC_STRING_ALTERNATIVE = TokenType.VAR
 615
 616    # Whether string escape characters function as such when placed within raw strings
 617    STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS = True
 618
 619    NESTED_COMMENTS = True
 620
 621    # Autofilled
 622    _COMMENTS: t.Dict[str, str] = {}
 623    _FORMAT_STRINGS: t.Dict[str, t.Tuple[str, TokenType]] = {}
 624    _IDENTIFIERS: t.Dict[str, str] = {}
 625    _IDENTIFIER_ESCAPES: t.Set[str] = set()
 626    _QUOTES: t.Dict[str, str] = {}
 627    _STRING_ESCAPES: t.Set[str] = set()
 628    _KEYWORD_TRIE: t.Dict = {}
 629    _RS_TOKENIZER: t.Optional[t.Any] = None
 630
 631    KEYWORDS: t.Dict[str, TokenType] = {
 632        **{f"{{%{postfix}": TokenType.BLOCK_START for postfix in ("", "+", "-")},
 633        **{f"{prefix}%}}": TokenType.BLOCK_END for prefix in ("", "+", "-")},
 634        **{f"{{{{{postfix}": TokenType.BLOCK_START for postfix in ("+", "-")},
 635        **{f"{prefix}}}}}": TokenType.BLOCK_END for prefix in ("+", "-")},
 636        "/*+": TokenType.HINT,
 637        "==": TokenType.EQ,
 638        "::": TokenType.DCOLON,
 639        "||": TokenType.DPIPE,
 640        ">=": TokenType.GTE,
 641        "<=": TokenType.LTE,
 642        "<>": TokenType.NEQ,
 643        "!=": TokenType.NEQ,
 644        ":=": TokenType.COLON_EQ,
 645        "<=>": TokenType.NULLSAFE_EQ,
 646        "->": TokenType.ARROW,
 647        "->>": TokenType.DARROW,
 648        "=>": TokenType.FARROW,
 649        "#>": TokenType.HASH_ARROW,
 650        "#>>": TokenType.DHASH_ARROW,
 651        "<->": TokenType.LR_ARROW,
 652        "&&": TokenType.DAMP,
 653        "??": TokenType.DQMARK,
 654        "~~~": TokenType.GLOB,
 655        "~~": TokenType.LIKE,
 656        "~~*": TokenType.ILIKE,
 657        "~*": TokenType.IRLIKE,
 658        "ALL": TokenType.ALL,
 659        "ALWAYS": TokenType.ALWAYS,
 660        "AND": TokenType.AND,
 661        "ANTI": TokenType.ANTI,
 662        "ANY": TokenType.ANY,
 663        "ASC": TokenType.ASC,
 664        "AS": TokenType.ALIAS,
 665        "ASOF": TokenType.ASOF,
 666        "AUTOINCREMENT": TokenType.AUTO_INCREMENT,
 667        "AUTO_INCREMENT": TokenType.AUTO_INCREMENT,
 668        "BEGIN": TokenType.BEGIN,
 669        "BETWEEN": TokenType.BETWEEN,
 670        "CACHE": TokenType.CACHE,
 671        "UNCACHE": TokenType.UNCACHE,
 672        "CASE": TokenType.CASE,
 673        "CHARACTER SET": TokenType.CHARACTER_SET,
 674        "CLUSTER BY": TokenType.CLUSTER_BY,
 675        "COLLATE": TokenType.COLLATE,
 676        "COLUMN": TokenType.COLUMN,
 677        "COMMIT": TokenType.COMMIT,
 678        "CONNECT BY": TokenType.CONNECT_BY,
 679        "CONSTRAINT": TokenType.CONSTRAINT,
 680        "COPY": TokenType.COPY,
 681        "CREATE": TokenType.CREATE,
 682        "CROSS": TokenType.CROSS,
 683        "CUBE": TokenType.CUBE,
 684        "CURRENT_DATE": TokenType.CURRENT_DATE,
 685        "CURRENT_TIME": TokenType.CURRENT_TIME,
 686        "CURRENT_TIMESTAMP": TokenType.CURRENT_TIMESTAMP,
 687        "CURRENT_USER": TokenType.CURRENT_USER,
 688        "DATABASE": TokenType.DATABASE,
 689        "DEFAULT": TokenType.DEFAULT,
 690        "DELETE": TokenType.DELETE,
 691        "DESC": TokenType.DESC,
 692        "DESCRIBE": TokenType.DESCRIBE,
 693        "DISTINCT": TokenType.DISTINCT,
 694        "DISTRIBUTE BY": TokenType.DISTRIBUTE_BY,
 695        "DIV": TokenType.DIV,
 696        "DROP": TokenType.DROP,
 697        "ELSE": TokenType.ELSE,
 698        "END": TokenType.END,
 699        "ENUM": TokenType.ENUM,
 700        "ESCAPE": TokenType.ESCAPE,
 701        "EXCEPT": TokenType.EXCEPT,
 702        "EXECUTE": TokenType.EXECUTE,
 703        "EXISTS": TokenType.EXISTS,
 704        "FALSE": TokenType.FALSE,
 705        "FETCH": TokenType.FETCH,
 706        "FILTER": TokenType.FILTER,
 707        "FIRST": TokenType.FIRST,
 708        "FULL": TokenType.FULL,
 709        "FUNCTION": TokenType.FUNCTION,
 710        "FOR": TokenType.FOR,
 711        "FOREIGN KEY": TokenType.FOREIGN_KEY,
 712        "FORMAT": TokenType.FORMAT,
 713        "FROM": TokenType.FROM,
 714        "GEOGRAPHY": TokenType.GEOGRAPHY,
 715        "GEOMETRY": TokenType.GEOMETRY,
 716        "GLOB": TokenType.GLOB,
 717        "GROUP BY": TokenType.GROUP_BY,
 718        "GROUPING SETS": TokenType.GROUPING_SETS,
 719        "HAVING": TokenType.HAVING,
 720        "ILIKE": TokenType.ILIKE,
 721        "IN": TokenType.IN,
 722        "INDEX": TokenType.INDEX,
 723        "INET": TokenType.INET,
 724        "INNER": TokenType.INNER,
 725        "INSERT": TokenType.INSERT,
 726        "INTERVAL": TokenType.INTERVAL,
 727        "INTERSECT": TokenType.INTERSECT,
 728        "INTO": TokenType.INTO,
 729        "IS": TokenType.IS,
 730        "ISNULL": TokenType.ISNULL,
 731        "JOIN": TokenType.JOIN,
 732        "KEEP": TokenType.KEEP,
 733        "KILL": TokenType.KILL,
 734        "LATERAL": TokenType.LATERAL,
 735        "LEFT": TokenType.LEFT,
 736        "LIKE": TokenType.LIKE,
 737        "LIMIT": TokenType.LIMIT,
 738        "LOAD": TokenType.LOAD,
 739        "LOCK": TokenType.LOCK,
 740        "MERGE": TokenType.MERGE,
 741        "NATURAL": TokenType.NATURAL,
 742        "NEXT": TokenType.NEXT,
 743        "NOT": TokenType.NOT,
 744        "NOTNULL": TokenType.NOTNULL,
 745        "NULL": TokenType.NULL,
 746        "OBJECT": TokenType.OBJECT,
 747        "OFFSET": TokenType.OFFSET,
 748        "ON": TokenType.ON,
 749        "OR": TokenType.OR,
 750        "XOR": TokenType.XOR,
 751        "ORDER BY": TokenType.ORDER_BY,
 752        "ORDINALITY": TokenType.ORDINALITY,
 753        "OUTER": TokenType.OUTER,
 754        "OVER": TokenType.OVER,
 755        "OVERLAPS": TokenType.OVERLAPS,
 756        "OVERWRITE": TokenType.OVERWRITE,
 757        "PARTITION": TokenType.PARTITION,
 758        "PARTITION BY": TokenType.PARTITION_BY,
 759        "PARTITIONED BY": TokenType.PARTITION_BY,
 760        "PARTITIONED_BY": TokenType.PARTITION_BY,
 761        "PERCENT": TokenType.PERCENT,
 762        "PIVOT": TokenType.PIVOT,
 763        "PRAGMA": TokenType.PRAGMA,
 764        "PRIMARY KEY": TokenType.PRIMARY_KEY,
 765        "PROCEDURE": TokenType.PROCEDURE,
 766        "QUALIFY": TokenType.QUALIFY,
 767        "RANGE": TokenType.RANGE,
 768        "RECURSIVE": TokenType.RECURSIVE,
 769        "REGEXP": TokenType.RLIKE,
 770        "RENAME": TokenType.RENAME,
 771        "REPLACE": TokenType.REPLACE,
 772        "RETURNING": TokenType.RETURNING,
 773        "REFERENCES": TokenType.REFERENCES,
 774        "RIGHT": TokenType.RIGHT,
 775        "RLIKE": TokenType.RLIKE,
 776        "ROLLBACK": TokenType.ROLLBACK,
 777        "ROLLUP": TokenType.ROLLUP,
 778        "ROW": TokenType.ROW,
 779        "ROWS": TokenType.ROWS,
 780        "SCHEMA": TokenType.SCHEMA,
 781        "SELECT": TokenType.SELECT,
 782        "SEMI": TokenType.SEMI,
 783        "SET": TokenType.SET,
 784        "SETTINGS": TokenType.SETTINGS,
 785        "SHOW": TokenType.SHOW,
 786        "SIMILAR TO": TokenType.SIMILAR_TO,
 787        "SOME": TokenType.SOME,
 788        "SORT BY": TokenType.SORT_BY,
 789        "START WITH": TokenType.START_WITH,
 790        "STRAIGHT_JOIN": TokenType.STRAIGHT_JOIN,
 791        "TABLE": TokenType.TABLE,
 792        "TABLESAMPLE": TokenType.TABLE_SAMPLE,
 793        "TEMP": TokenType.TEMPORARY,
 794        "TEMPORARY": TokenType.TEMPORARY,
 795        "THEN": TokenType.THEN,
 796        "TRUE": TokenType.TRUE,
 797        "TRUNCATE": TokenType.TRUNCATE,
 798        "UNION": TokenType.UNION,
 799        "UNKNOWN": TokenType.UNKNOWN,
 800        "UNNEST": TokenType.UNNEST,
 801        "UNPIVOT": TokenType.UNPIVOT,
 802        "UPDATE": TokenType.UPDATE,
 803        "USE": TokenType.USE,
 804        "USING": TokenType.USING,
 805        "UUID": TokenType.UUID,
 806        "VALUES": TokenType.VALUES,
 807        "VIEW": TokenType.VIEW,
 808        "VOLATILE": TokenType.VOLATILE,
 809        "WHEN": TokenType.WHEN,
 810        "WHERE": TokenType.WHERE,
 811        "WINDOW": TokenType.WINDOW,
 812        "WITH": TokenType.WITH,
 813        "APPLY": TokenType.APPLY,
 814        "ARRAY": TokenType.ARRAY,
 815        "BIT": TokenType.BIT,
 816        "BOOL": TokenType.BOOLEAN,
 817        "BOOLEAN": TokenType.BOOLEAN,
 818        "BYTE": TokenType.TINYINT,
 819        "MEDIUMINT": TokenType.MEDIUMINT,
 820        "INT1": TokenType.TINYINT,
 821        "TINYINT": TokenType.TINYINT,
 822        "INT16": TokenType.SMALLINT,
 823        "SHORT": TokenType.SMALLINT,
 824        "SMALLINT": TokenType.SMALLINT,
 825        "INT128": TokenType.INT128,
 826        "HUGEINT": TokenType.INT128,
 827        "UHUGEINT": TokenType.UINT128,
 828        "INT2": TokenType.SMALLINT,
 829        "INTEGER": TokenType.INT,
 830        "INT": TokenType.INT,
 831        "INT4": TokenType.INT,
 832        "INT32": TokenType.INT,
 833        "INT64": TokenType.BIGINT,
 834        "LONG": TokenType.BIGINT,
 835        "BIGINT": TokenType.BIGINT,
 836        "INT8": TokenType.TINYINT,
 837        "UINT": TokenType.UINT,
 838        "DEC": TokenType.DECIMAL,
 839        "DECIMAL": TokenType.DECIMAL,
 840        "DECIMAL32": TokenType.DECIMAL32,
 841        "DECIMAL64": TokenType.DECIMAL64,
 842        "DECIMAL128": TokenType.DECIMAL128,
 843        "BIGDECIMAL": TokenType.BIGDECIMAL,
 844        "BIGNUMERIC": TokenType.BIGDECIMAL,
 845        "LIST": TokenType.LIST,
 846        "MAP": TokenType.MAP,
 847        "NULLABLE": TokenType.NULLABLE,
 848        "NUMBER": TokenType.DECIMAL,
 849        "NUMERIC": TokenType.DECIMAL,
 850        "FIXED": TokenType.DECIMAL,
 851        "REAL": TokenType.FLOAT,
 852        "FLOAT": TokenType.FLOAT,
 853        "FLOAT4": TokenType.FLOAT,
 854        "FLOAT8": TokenType.DOUBLE,
 855        "DOUBLE": TokenType.DOUBLE,
 856        "DOUBLE PRECISION": TokenType.DOUBLE,
 857        "JSON": TokenType.JSON,
 858        "JSONB": TokenType.JSONB,
 859        "CHAR": TokenType.CHAR,
 860        "CHARACTER": TokenType.CHAR,
 861        "NCHAR": TokenType.NCHAR,
 862        "VARCHAR": TokenType.VARCHAR,
 863        "VARCHAR2": TokenType.VARCHAR,
 864        "NVARCHAR": TokenType.NVARCHAR,
 865        "NVARCHAR2": TokenType.NVARCHAR,
 866        "BPCHAR": TokenType.BPCHAR,
 867        "STR": TokenType.TEXT,
 868        "STRING": TokenType.TEXT,
 869        "TEXT": TokenType.TEXT,
 870        "LONGTEXT": TokenType.LONGTEXT,
 871        "MEDIUMTEXT": TokenType.MEDIUMTEXT,
 872        "TINYTEXT": TokenType.TINYTEXT,
 873        "CLOB": TokenType.TEXT,
 874        "LONGVARCHAR": TokenType.TEXT,
 875        "BINARY": TokenType.BINARY,
 876        "BLOB": TokenType.VARBINARY,
 877        "LONGBLOB": TokenType.LONGBLOB,
 878        "MEDIUMBLOB": TokenType.MEDIUMBLOB,
 879        "TINYBLOB": TokenType.TINYBLOB,
 880        "BYTEA": TokenType.VARBINARY,
 881        "VARBINARY": TokenType.VARBINARY,
 882        "TIME": TokenType.TIME,
 883        "TIMETZ": TokenType.TIMETZ,
 884        "TIMESTAMP": TokenType.TIMESTAMP,
 885        "TIMESTAMPTZ": TokenType.TIMESTAMPTZ,
 886        "TIMESTAMPLTZ": TokenType.TIMESTAMPLTZ,
 887        "TIMESTAMP_LTZ": TokenType.TIMESTAMPLTZ,
 888        "TIMESTAMPNTZ": TokenType.TIMESTAMPNTZ,
 889        "TIMESTAMP_NTZ": TokenType.TIMESTAMPNTZ,
 890        "DATE": TokenType.DATE,
 891        "DATETIME": TokenType.DATETIME,
 892        "INT4RANGE": TokenType.INT4RANGE,
 893        "INT4MULTIRANGE": TokenType.INT4MULTIRANGE,
 894        "INT8RANGE": TokenType.INT8RANGE,
 895        "INT8MULTIRANGE": TokenType.INT8MULTIRANGE,
 896        "NUMRANGE": TokenType.NUMRANGE,
 897        "NUMMULTIRANGE": TokenType.NUMMULTIRANGE,
 898        "TSRANGE": TokenType.TSRANGE,
 899        "TSMULTIRANGE": TokenType.TSMULTIRANGE,
 900        "TSTZRANGE": TokenType.TSTZRANGE,
 901        "TSTZMULTIRANGE": TokenType.TSTZMULTIRANGE,
 902        "DATERANGE": TokenType.DATERANGE,
 903        "DATEMULTIRANGE": TokenType.DATEMULTIRANGE,
 904        "UNIQUE": TokenType.UNIQUE,
 905        "VECTOR": TokenType.VECTOR,
 906        "STRUCT": TokenType.STRUCT,
 907        "SEQUENCE": TokenType.SEQUENCE,
 908        "VARIANT": TokenType.VARIANT,
 909        "ALTER": TokenType.ALTER,
 910        "ANALYZE": TokenType.COMMAND,
 911        "CALL": TokenType.COMMAND,
 912        "COMMENT": TokenType.COMMENT,
 913        "EXPLAIN": TokenType.COMMAND,
 914        "GRANT": TokenType.GRANT,
 915        "OPTIMIZE": TokenType.COMMAND,
 916        "PREPARE": TokenType.COMMAND,
 917        "VACUUM": TokenType.COMMAND,
 918        "USER-DEFINED": TokenType.USERDEFINED,
 919        "FOR VERSION": TokenType.VERSION_SNAPSHOT,
 920        "FOR TIMESTAMP": TokenType.TIMESTAMP_SNAPSHOT,
 921    }
 922
 923    WHITE_SPACE: t.Dict[t.Optional[str], TokenType] = {
 924        " ": TokenType.SPACE,
 925        "\t": TokenType.SPACE,
 926        "\n": TokenType.BREAK,
 927        "\r": TokenType.BREAK,
 928    }
 929
 930    COMMANDS = {
 931        TokenType.COMMAND,
 932        TokenType.EXECUTE,
 933        TokenType.FETCH,
 934        TokenType.SHOW,
 935        TokenType.RENAME,
 936    }
 937
 938    COMMAND_PREFIX_TOKENS = {TokenType.SEMICOLON, TokenType.BEGIN}
 939
 940    # Handle numeric literals like in hive (3L = BIGINT)
 941    NUMERIC_LITERALS: t.Dict[str, str] = {}
 942
 943    COMMENTS = ["--", ("/*", "*/")]
 944
 945    __slots__ = (
 946        "sql",
 947        "size",
 948        "tokens",
 949        "dialect",
 950        "_start",
 951        "_current",
 952        "_line",
 953        "_col",
 954        "_comments",
 955        "_char",
 956        "_end",
 957        "_peek",
 958        "_prev_token_line",
 959        "_rs_dialect_settings",
 960    )
 961
 962    def __init__(self, dialect: DialectType = None) -> None:
 963        from sqlglot.dialects import Dialect
 964
 965        self.dialect = Dialect.get_or_raise(dialect)
 966
 967        if USE_RS_TOKENIZER:
 968            self._rs_dialect_settings = RsTokenizerDialectSettings(
 969                unescaped_sequences=self.dialect.UNESCAPED_SEQUENCES,
 970                identifiers_can_start_with_digit=self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT,
 971            )
 972
 973        self.reset()
 974
 975    def reset(self) -> None:
 976        self.sql = ""
 977        self.size = 0
 978        self.tokens: t.List[Token] = []
 979        self._start = 0
 980        self._current = 0
 981        self._line = 1
 982        self._col = 0
 983        self._comments: t.List[str] = []
 984
 985        self._char = ""
 986        self._end = False
 987        self._peek = ""
 988        self._prev_token_line = -1
 989
 990    def tokenize(self, sql: str) -> t.List[Token]:
 991        """Returns a list of tokens corresponding to the SQL string `sql`."""
 992        if USE_RS_TOKENIZER:
 993            return self.tokenize_rs(sql)
 994
 995        self.reset()
 996        self.sql = sql
 997        self.size = len(sql)
 998
 999        try:
1000            self._scan()
1001        except Exception as e:
1002            start = max(self._current - 50, 0)
1003            end = min(self._current + 50, self.size - 1)
1004            context = self.sql[start:end]
1005            raise TokenError(f"Error tokenizing '{context}'") from e
1006
1007        return self.tokens
1008
1009    def _scan(self, until: t.Optional[t.Callable] = None) -> None:
1010        while self.size and not self._end:
1011            current = self._current
1012
1013            # Skip spaces here rather than iteratively calling advance() for performance reasons
1014            while current < self.size:
1015                char = self.sql[current]
1016
1017                if char.isspace() and (char == " " or char == "\t"):
1018                    current += 1
1019                else:
1020                    break
1021
1022            offset = current - self._current if current > self._current else 1
1023
1024            self._start = current
1025            self._advance(offset)
1026
1027            if not self._char.isspace():
1028                if self._char.isdigit():
1029                    self._scan_number()
1030                elif self._char in self._IDENTIFIERS:
1031                    self._scan_identifier(self._IDENTIFIERS[self._char])
1032                else:
1033                    self._scan_keywords()
1034
1035            if until and until():
1036                break
1037
1038        if self.tokens and self._comments:
1039            self.tokens[-1].comments.extend(self._comments)
1040
1041    def _chars(self, size: int) -> str:
1042        if size == 1:
1043            return self._char
1044
1045        start = self._current - 1
1046        end = start + size
1047
1048        return self.sql[start:end] if end <= self.size else ""
1049
1050    def _advance(self, i: int = 1, alnum: bool = False) -> None:
1051        if self.WHITE_SPACE.get(self._char) is TokenType.BREAK:
1052            # Ensures we don't count an extra line if we get a \r\n line break sequence
1053            if not (self._char == "\r" and self._peek == "\n"):
1054                self._col = 1
1055                self._line += 1
1056        else:
1057            self._col += i
1058
1059        self._current += i
1060        self._end = self._current >= self.size
1061        self._char = self.sql[self._current - 1]
1062        self._peek = "" if self._end else self.sql[self._current]
1063
1064        if alnum and self._char.isalnum():
1065            # Here we use local variables instead of attributes for better performance
1066            _col = self._col
1067            _current = self._current
1068            _end = self._end
1069            _peek = self._peek
1070
1071            while _peek.isalnum():
1072                _col += 1
1073                _current += 1
1074                _end = _current >= self.size
1075                _peek = "" if _end else self.sql[_current]
1076
1077            self._col = _col
1078            self._current = _current
1079            self._end = _end
1080            self._peek = _peek
1081            self._char = self.sql[_current - 1]
1082
1083    @property
1084    def _text(self) -> str:
1085        return self.sql[self._start : self._current]
1086
1087    def _add(self, token_type: TokenType, text: t.Optional[str] = None) -> None:
1088        self._prev_token_line = self._line
1089
1090        if self._comments and token_type == TokenType.SEMICOLON and self.tokens:
1091            self.tokens[-1].comments.extend(self._comments)
1092            self._comments = []
1093
1094        self.tokens.append(
1095            Token(
1096                token_type,
1097                text=self._text if text is None else text,
1098                line=self._line,
1099                col=self._col,
1100                start=self._start,
1101                end=self._current - 1,
1102                comments=self._comments,
1103            )
1104        )
1105        self._comments = []
1106
1107        # If we have either a semicolon or a begin token before the command's token, we'll parse
1108        # whatever follows the command's token as a string
1109        if (
1110            token_type in self.COMMANDS
1111            and self._peek != ";"
1112            and (len(self.tokens) == 1 or self.tokens[-2].token_type in self.COMMAND_PREFIX_TOKENS)
1113        ):
1114            start = self._current
1115            tokens = len(self.tokens)
1116            self._scan(lambda: self._peek == ";")
1117            self.tokens = self.tokens[:tokens]
1118            text = self.sql[start : self._current].strip()
1119            if text:
1120                self._add(TokenType.STRING, text)
1121
1122    def _scan_keywords(self) -> None:
1123        size = 0
1124        word = None
1125        chars = self._text
1126        char = chars
1127        prev_space = False
1128        skip = False
1129        trie = self._KEYWORD_TRIE
1130        single_token = char in self.SINGLE_TOKENS
1131
1132        while chars:
1133            if skip:
1134                result = TrieResult.PREFIX
1135            else:
1136                result, trie = in_trie(trie, char.upper())
1137
1138            if result == TrieResult.FAILED:
1139                break
1140            if result == TrieResult.EXISTS:
1141                word = chars
1142
1143            end = self._current + size
1144            size += 1
1145
1146            if end < self.size:
1147                char = self.sql[end]
1148                single_token = single_token or char in self.SINGLE_TOKENS
1149                is_space = char.isspace()
1150
1151                if not is_space or not prev_space:
1152                    if is_space:
1153                        char = " "
1154                    chars += char
1155                    prev_space = is_space
1156                    skip = False
1157                else:
1158                    skip = True
1159            else:
1160                char = ""
1161                break
1162
1163        if word:
1164            if self._scan_string(word):
1165                return
1166            if self._scan_comment(word):
1167                return
1168            if prev_space or single_token or not char:
1169                self._advance(size - 1)
1170                word = word.upper()
1171                self._add(self.KEYWORDS[word], text=word)
1172                return
1173
1174        if self._char in self.SINGLE_TOKENS:
1175            self._add(self.SINGLE_TOKENS[self._char], text=self._char)
1176            return
1177
1178        self._scan_var()
1179
1180    def _scan_comment(self, comment_start: str) -> bool:
1181        if comment_start not in self._COMMENTS:
1182            return False
1183
1184        comment_start_line = self._line
1185        comment_start_size = len(comment_start)
1186        comment_end = self._COMMENTS[comment_start]
1187
1188        if comment_end:
1189            # Skip the comment's start delimiter
1190            self._advance(comment_start_size)
1191
1192            comment_count = 1
1193            comment_end_size = len(comment_end)
1194
1195            while not self._end:
1196                if self._chars(comment_end_size) == comment_end:
1197                    comment_count -= 1
1198                    if not comment_count:
1199                        break
1200
1201                self._advance(alnum=True)
1202
1203                # Nested comments are allowed by some dialects, e.g. databricks, duckdb, postgres
1204                if (
1205                    self.NESTED_COMMENTS
1206                    and not self._end
1207                    and self._chars(comment_end_size) == comment_start
1208                ):
1209                    self._advance(comment_start_size)
1210                    comment_count += 1
1211
1212            self._comments.append(self._text[comment_start_size : -comment_end_size + 1])
1213            self._advance(comment_end_size - 1)
1214        else:
1215            while not self._end and self.WHITE_SPACE.get(self._peek) is not TokenType.BREAK:
1216                self._advance(alnum=True)
1217            self._comments.append(self._text[comment_start_size:])
1218
1219        # Leading comment is attached to the succeeding token, whilst trailing comment to the preceding.
1220        # Multiple consecutive comments are preserved by appending them to the current comments list.
1221        if comment_start_line == self._prev_token_line:
1222            self.tokens[-1].comments.extend(self._comments)
1223            self._comments = []
1224            self._prev_token_line = self._line
1225
1226        return True
1227
1228    def _scan_number(self) -> None:
1229        if self._char == "0":
1230            peek = self._peek.upper()
1231            if peek == "B":
1232                return self._scan_bits() if self.BIT_STRINGS else self._add(TokenType.NUMBER)
1233            elif peek == "X":
1234                return self._scan_hex() if self.HEX_STRINGS else self._add(TokenType.NUMBER)
1235
1236        decimal = False
1237        scientific = 0
1238
1239        while True:
1240            if self._peek.isdigit():
1241                self._advance()
1242            elif self._peek == "." and not decimal:
1243                if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER:
1244                    return self._add(TokenType.NUMBER)
1245                decimal = True
1246                self._advance()
1247            elif self._peek in ("-", "+") and scientific == 1:
1248                scientific += 1
1249                self._advance()
1250            elif self._peek.upper() == "E" and not scientific:
1251                scientific += 1
1252                self._advance()
1253            elif self._peek.isidentifier():
1254                number_text = self._text
1255                literal = ""
1256
1257                while self._peek.strip() and self._peek not in self.SINGLE_TOKENS:
1258                    literal += self._peek
1259                    self._advance()
1260
1261                token_type = self.KEYWORDS.get(self.NUMERIC_LITERALS.get(literal.upper(), ""))
1262
1263                if token_type:
1264                    self._add(TokenType.NUMBER, number_text)
1265                    self._add(TokenType.DCOLON, "::")
1266                    return self._add(token_type, literal)
1267                elif self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT:
1268                    return self._add(TokenType.VAR)
1269
1270                self._advance(-len(literal))
1271                return self._add(TokenType.NUMBER, number_text)
1272            else:
1273                return self._add(TokenType.NUMBER)
1274
1275    def _scan_bits(self) -> None:
1276        self._advance()
1277        value = self._extract_value()
1278        try:
1279            # If `value` can't be converted to a binary, fallback to tokenizing it as an identifier
1280            int(value, 2)
1281            self._add(TokenType.BIT_STRING, value[2:])  # Drop the 0b
1282        except ValueError:
1283            self._add(TokenType.IDENTIFIER)
1284
1285    def _scan_hex(self) -> None:
1286        self._advance()
1287        value = self._extract_value()
1288        try:
1289            # If `value` can't be converted to a hex, fallback to tokenizing it as an identifier
1290            int(value, 16)
1291            self._add(TokenType.HEX_STRING, value[2:])  # Drop the 0x
1292        except ValueError:
1293            self._add(TokenType.IDENTIFIER)
1294
1295    def _extract_value(self) -> str:
1296        while True:
1297            char = self._peek.strip()
1298            if char and char not in self.SINGLE_TOKENS:
1299                self._advance(alnum=True)
1300            else:
1301                break
1302
1303        return self._text
1304
1305    def _scan_string(self, start: str) -> bool:
1306        base = None
1307        token_type = TokenType.STRING
1308
1309        if start in self._QUOTES:
1310            end = self._QUOTES[start]
1311        elif start in self._FORMAT_STRINGS:
1312            end, token_type = self._FORMAT_STRINGS[start]
1313
1314            if token_type == TokenType.HEX_STRING:
1315                base = 16
1316            elif token_type == TokenType.BIT_STRING:
1317                base = 2
1318            elif token_type == TokenType.HEREDOC_STRING:
1319                self._advance()
1320
1321                if self._char == end:
1322                    tag = ""
1323                else:
1324                    tag = self._extract_string(
1325                        end,
1326                        raw_string=True,
1327                        raise_unmatched=not self.HEREDOC_TAG_IS_IDENTIFIER,
1328                    )
1329
1330                if tag and self.HEREDOC_TAG_IS_IDENTIFIER and (self._end or not tag.isidentifier()):
1331                    if not self._end:
1332                        self._advance(-1)
1333
1334                    self._advance(-len(tag))
1335                    self._add(self.HEREDOC_STRING_ALTERNATIVE)
1336                    return True
1337
1338                end = f"{start}{tag}{end}"
1339        else:
1340            return False
1341
1342        self._advance(len(start))
1343        text = self._extract_string(end, raw_string=token_type == TokenType.RAW_STRING)
1344
1345        if base:
1346            try:
1347                int(text, base)
1348            except Exception:
1349                raise TokenError(
1350                    f"Numeric string contains invalid characters from {self._line}:{self._start}"
1351                )
1352
1353        self._add(token_type, text)
1354        return True
1355
1356    def _scan_identifier(self, identifier_end: str) -> None:
1357        self._advance()
1358        text = self._extract_string(identifier_end, escapes=self._IDENTIFIER_ESCAPES)
1359        self._add(TokenType.IDENTIFIER, text)
1360
1361    def _scan_var(self) -> None:
1362        while True:
1363            char = self._peek.strip()
1364            if char and (char in self.VAR_SINGLE_TOKENS or char not in self.SINGLE_TOKENS):
1365                self._advance(alnum=True)
1366            else:
1367                break
1368
1369        self._add(
1370            TokenType.VAR
1371            if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER
1372            else self.KEYWORDS.get(self._text.upper(), TokenType.VAR)
1373        )
1374
1375    def _extract_string(
1376        self,
1377        delimiter: str,
1378        escapes: t.Optional[t.Set[str]] = None,
1379        raw_string: bool = False,
1380        raise_unmatched: bool = True,
1381    ) -> str:
1382        text = ""
1383        delim_size = len(delimiter)
1384        escapes = self._STRING_ESCAPES if escapes is None else escapes
1385
1386        while True:
1387            if (
1388                not raw_string
1389                and self.dialect.UNESCAPED_SEQUENCES
1390                and self._peek
1391                and self._char in self.STRING_ESCAPES
1392            ):
1393                unescaped_sequence = self.dialect.UNESCAPED_SEQUENCES.get(self._char + self._peek)
1394                if unescaped_sequence:
1395                    self._advance(2)
1396                    text += unescaped_sequence
1397                    continue
1398            if (
1399                (self.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS or not raw_string)
1400                and self._char in escapes
1401                and (self._peek == delimiter or self._peek in escapes)
1402                and (self._char not in self._QUOTES or self._char == self._peek)
1403            ):
1404                if self._peek == delimiter:
1405                    text += self._peek
1406                else:
1407                    text += self._char + self._peek
1408
1409                if self._current + 1 < self.size:
1410                    self._advance(2)
1411                else:
1412                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._current}")
1413            else:
1414                if self._chars(delim_size) == delimiter:
1415                    if delim_size > 1:
1416                        self._advance(delim_size - 1)
1417                    break
1418
1419                if self._end:
1420                    if not raise_unmatched:
1421                        return text + self._char
1422
1423                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._start}")
1424
1425                current = self._current - 1
1426                self._advance(alnum=True)
1427                text += self.sql[current : self._current - 1]
1428
1429        return text
1430
1431    def tokenize_rs(self, sql: str) -> t.List[Token]:
1432        if not self._RS_TOKENIZER:
1433            raise SqlglotError("Rust tokenizer is not available")
1434
1435        try:
1436            tokens = self._RS_TOKENIZER.tokenize(sql, self._rs_dialect_settings)
1437            for token in tokens:
1438                token.token_type = _ALL_TOKEN_TYPES[token.token_type_index]
1439            return tokens
1440        except Exception as e:
1441            raise TokenError(str(e))