sqlglot.tokens

   1from __future__ import annotations
   2
   3import os
   4import typing as t
   5from enum import auto
   6
   7from sqlglot.errors import SqlglotError, TokenError
   8from sqlglot.helper import AutoName
   9from sqlglot.trie import TrieResult, in_trie, new_trie
  10
  11if t.TYPE_CHECKING:
  12    from sqlglot.dialects.dialect import DialectType
  13
  14
  15try:
  16    from sqlglotrs import (  # type: ignore
  17        Tokenizer as RsTokenizer,
  18        TokenizerDialectSettings as RsTokenizerDialectSettings,
  19        TokenizerSettings as RsTokenizerSettings,
  20        TokenTypeSettings as RsTokenTypeSettings,
  21    )
  22
  23    USE_RS_TOKENIZER = os.environ.get("SQLGLOTRS_TOKENIZER", "1") == "1"
  24except ImportError:
  25    USE_RS_TOKENIZER = False
  26
  27
  28class TokenType(AutoName):
  29    L_PAREN = auto()
  30    R_PAREN = auto()
  31    L_BRACKET = auto()
  32    R_BRACKET = auto()
  33    L_BRACE = auto()
  34    R_BRACE = auto()
  35    COMMA = auto()
  36    DOT = auto()
  37    DASH = auto()
  38    PLUS = auto()
  39    COLON = auto()
  40    DOTCOLON = auto()
  41    DCOLON = auto()
  42    DQMARK = auto()
  43    SEMICOLON = auto()
  44    STAR = auto()
  45    BACKSLASH = auto()
  46    SLASH = auto()
  47    LT = auto()
  48    LTE = auto()
  49    GT = auto()
  50    GTE = auto()
  51    NOT = auto()
  52    EQ = auto()
  53    NEQ = auto()
  54    NULLSAFE_EQ = auto()
  55    COLON_EQ = auto()
  56    AND = auto()
  57    OR = auto()
  58    AMP = auto()
  59    DPIPE = auto()
  60    PIPE_GT = auto()
  61    PIPE = auto()
  62    PIPE_SLASH = auto()
  63    DPIPE_SLASH = auto()
  64    CARET = auto()
  65    CARET_AT = auto()
  66    TILDA = auto()
  67    ARROW = auto()
  68    DARROW = auto()
  69    FARROW = auto()
  70    HASH = auto()
  71    HASH_ARROW = auto()
  72    DHASH_ARROW = auto()
  73    LR_ARROW = auto()
  74    DAT = auto()
  75    LT_AT = auto()
  76    AT_GT = auto()
  77    DOLLAR = auto()
  78    PARAMETER = auto()
  79    SESSION_PARAMETER = auto()
  80    DAMP = auto()
  81    XOR = auto()
  82    DSTAR = auto()
  83
  84    URI_START = auto()
  85
  86    BLOCK_START = auto()
  87    BLOCK_END = auto()
  88
  89    SPACE = auto()
  90    BREAK = auto()
  91
  92    STRING = auto()
  93    NUMBER = auto()
  94    IDENTIFIER = auto()
  95    DATABASE = auto()
  96    COLUMN = auto()
  97    COLUMN_DEF = auto()
  98    SCHEMA = auto()
  99    TABLE = auto()
 100    WAREHOUSE = auto()
 101    STAGE = auto()
 102    STREAMLIT = auto()
 103    VAR = auto()
 104    BIT_STRING = auto()
 105    HEX_STRING = auto()
 106    BYTE_STRING = auto()
 107    NATIONAL_STRING = auto()
 108    RAW_STRING = auto()
 109    HEREDOC_STRING = auto()
 110    UNICODE_STRING = auto()
 111
 112    # types
 113    BIT = auto()
 114    BOOLEAN = auto()
 115    TINYINT = auto()
 116    UTINYINT = auto()
 117    SMALLINT = auto()
 118    USMALLINT = auto()
 119    MEDIUMINT = auto()
 120    UMEDIUMINT = auto()
 121    INT = auto()
 122    UINT = auto()
 123    BIGINT = auto()
 124    UBIGINT = auto()
 125    INT128 = auto()
 126    UINT128 = auto()
 127    INT256 = auto()
 128    UINT256 = auto()
 129    FLOAT = auto()
 130    DOUBLE = auto()
 131    UDOUBLE = auto()
 132    DECIMAL = auto()
 133    DECIMAL32 = auto()
 134    DECIMAL64 = auto()
 135    DECIMAL128 = auto()
 136    DECIMAL256 = auto()
 137    UDECIMAL = auto()
 138    BIGDECIMAL = auto()
 139    CHAR = auto()
 140    NCHAR = auto()
 141    VARCHAR = auto()
 142    NVARCHAR = auto()
 143    BPCHAR = auto()
 144    TEXT = auto()
 145    MEDIUMTEXT = auto()
 146    LONGTEXT = auto()
 147    BLOB = auto()
 148    MEDIUMBLOB = auto()
 149    LONGBLOB = auto()
 150    TINYBLOB = auto()
 151    TINYTEXT = auto()
 152    NAME = auto()
 153    BINARY = auto()
 154    VARBINARY = auto()
 155    JSON = auto()
 156    JSONB = auto()
 157    TIME = auto()
 158    TIMETZ = auto()
 159    TIMESTAMP = auto()
 160    TIMESTAMPTZ = auto()
 161    TIMESTAMPLTZ = auto()
 162    TIMESTAMPNTZ = auto()
 163    TIMESTAMP_S = auto()
 164    TIMESTAMP_MS = auto()
 165    TIMESTAMP_NS = auto()
 166    DATETIME = auto()
 167    DATETIME2 = auto()
 168    DATETIME64 = auto()
 169    SMALLDATETIME = auto()
 170    DATE = auto()
 171    DATE32 = auto()
 172    INT4RANGE = auto()
 173    INT4MULTIRANGE = auto()
 174    INT8RANGE = auto()
 175    INT8MULTIRANGE = auto()
 176    NUMRANGE = auto()
 177    NUMMULTIRANGE = auto()
 178    TSRANGE = auto()
 179    TSMULTIRANGE = auto()
 180    TSTZRANGE = auto()
 181    TSTZMULTIRANGE = auto()
 182    DATERANGE = auto()
 183    DATEMULTIRANGE = auto()
 184    UUID = auto()
 185    GEOGRAPHY = auto()
 186    NULLABLE = auto()
 187    GEOMETRY = auto()
 188    POINT = auto()
 189    RING = auto()
 190    LINESTRING = auto()
 191    MULTILINESTRING = auto()
 192    POLYGON = auto()
 193    MULTIPOLYGON = auto()
 194    HLLSKETCH = auto()
 195    HSTORE = auto()
 196    SUPER = auto()
 197    SERIAL = auto()
 198    SMALLSERIAL = auto()
 199    BIGSERIAL = auto()
 200    XML = auto()
 201    YEAR = auto()
 202    USERDEFINED = auto()
 203    MONEY = auto()
 204    SMALLMONEY = auto()
 205    ROWVERSION = auto()
 206    IMAGE = auto()
 207    VARIANT = auto()
 208    OBJECT = auto()
 209    INET = auto()
 210    IPADDRESS = auto()
 211    IPPREFIX = auto()
 212    IPV4 = auto()
 213    IPV6 = auto()
 214    ENUM = auto()
 215    ENUM8 = auto()
 216    ENUM16 = auto()
 217    FIXEDSTRING = auto()
 218    LOWCARDINALITY = auto()
 219    NESTED = auto()
 220    AGGREGATEFUNCTION = auto()
 221    SIMPLEAGGREGATEFUNCTION = auto()
 222    TDIGEST = auto()
 223    UNKNOWN = auto()
 224    VECTOR = auto()
 225    DYNAMIC = auto()
 226    VOID = auto()
 227
 228    # keywords
 229    ALIAS = auto()
 230    ALTER = auto()
 231    ALWAYS = auto()
 232    ALL = auto()
 233    ANTI = auto()
 234    ANY = auto()
 235    APPLY = auto()
 236    ARRAY = auto()
 237    ASC = auto()
 238    ASOF = auto()
 239    ATTACH = auto()
 240    AUTO_INCREMENT = auto()
 241    BEGIN = auto()
 242    BETWEEN = auto()
 243    BULK_COLLECT_INTO = auto()
 244    CACHE = auto()
 245    CASE = auto()
 246    CHARACTER_SET = auto()
 247    CLUSTER_BY = auto()
 248    COLLATE = auto()
 249    COMMAND = auto()
 250    COMMENT = auto()
 251    COMMIT = auto()
 252    CONNECT_BY = auto()
 253    CONSTRAINT = auto()
 254    COPY = auto()
 255    CREATE = auto()
 256    CROSS = auto()
 257    CUBE = auto()
 258    CURRENT_DATE = auto()
 259    CURRENT_DATETIME = auto()
 260    CURRENT_SCHEMA = auto()
 261    CURRENT_TIME = auto()
 262    CURRENT_TIMESTAMP = auto()
 263    CURRENT_USER = auto()
 264    DECLARE = auto()
 265    DEFAULT = auto()
 266    DELETE = auto()
 267    DESC = auto()
 268    DESCRIBE = auto()
 269    DETACH = auto()
 270    DICTIONARY = auto()
 271    DISTINCT = auto()
 272    DISTRIBUTE_BY = auto()
 273    DIV = auto()
 274    DROP = auto()
 275    ELSE = auto()
 276    END = auto()
 277    ESCAPE = auto()
 278    EXCEPT = auto()
 279    EXECUTE = auto()
 280    EXISTS = auto()
 281    FALSE = auto()
 282    FETCH = auto()
 283    FILE_FORMAT = auto()
 284    FILTER = auto()
 285    FINAL = auto()
 286    FIRST = auto()
 287    FOR = auto()
 288    FORCE = auto()
 289    FOREIGN_KEY = auto()
 290    FORMAT = auto()
 291    FROM = auto()
 292    FULL = auto()
 293    FUNCTION = auto()
 294    GET = auto()
 295    GLOB = auto()
 296    GLOBAL = auto()
 297    GRANT = auto()
 298    GROUP_BY = auto()
 299    GROUPING_SETS = auto()
 300    HAVING = auto()
 301    HINT = auto()
 302    IGNORE = auto()
 303    ILIKE = auto()
 304    ILIKE_ANY = auto()
 305    IN = auto()
 306    INDEX = auto()
 307    INNER = auto()
 308    INSERT = auto()
 309    INTERSECT = auto()
 310    INTERVAL = auto()
 311    INTO = auto()
 312    INTRODUCER = auto()
 313    IRLIKE = auto()
 314    IS = auto()
 315    ISNULL = auto()
 316    JOIN = auto()
 317    JOIN_MARKER = auto()
 318    KEEP = auto()
 319    KEY = auto()
 320    KILL = auto()
 321    LANGUAGE = auto()
 322    LATERAL = auto()
 323    LEFT = auto()
 324    LIKE = auto()
 325    LIKE_ANY = auto()
 326    LIMIT = auto()
 327    LIST = auto()
 328    LOAD = auto()
 329    LOCK = auto()
 330    MAP = auto()
 331    MATCH_CONDITION = auto()
 332    MATCH_RECOGNIZE = auto()
 333    MEMBER_OF = auto()
 334    MERGE = auto()
 335    MOD = auto()
 336    MODEL = auto()
 337    NATURAL = auto()
 338    NEXT = auto()
 339    NOTHING = auto()
 340    NOTNULL = auto()
 341    NULL = auto()
 342    OBJECT_IDENTIFIER = auto()
 343    OFFSET = auto()
 344    ON = auto()
 345    ONLY = auto()
 346    OPERATOR = auto()
 347    ORDER_BY = auto()
 348    ORDER_SIBLINGS_BY = auto()
 349    ORDERED = auto()
 350    ORDINALITY = auto()
 351    OUTER = auto()
 352    OVER = auto()
 353    OVERLAPS = auto()
 354    OVERWRITE = auto()
 355    PARTITION = auto()
 356    PARTITION_BY = auto()
 357    PERCENT = auto()
 358    PIVOT = auto()
 359    PLACEHOLDER = auto()
 360    POSITIONAL = auto()
 361    PRAGMA = auto()
 362    PREWHERE = auto()
 363    PRIMARY_KEY = auto()
 364    PROCEDURE = auto()
 365    PROPERTIES = auto()
 366    PSEUDO_TYPE = auto()
 367    PUT = auto()
 368    QUALIFY = auto()
 369    QUOTE = auto()
 370    RANGE = auto()
 371    RECURSIVE = auto()
 372    REFRESH = auto()
 373    RENAME = auto()
 374    REPLACE = auto()
 375    RETURNING = auto()
 376    REFERENCES = auto()
 377    RIGHT = auto()
 378    RLIKE = auto()
 379    ROLLBACK = auto()
 380    ROLLUP = auto()
 381    ROW = auto()
 382    ROWS = auto()
 383    SELECT = auto()
 384    SEMI = auto()
 385    SEPARATOR = auto()
 386    SEQUENCE = auto()
 387    SERDE_PROPERTIES = auto()
 388    SET = auto()
 389    SETTINGS = auto()
 390    SHOW = auto()
 391    SIMILAR_TO = auto()
 392    SOME = auto()
 393    SORT_BY = auto()
 394    START_WITH = auto()
 395    STORAGE_INTEGRATION = auto()
 396    STRAIGHT_JOIN = auto()
 397    STRUCT = auto()
 398    SUMMARIZE = auto()
 399    TABLE_SAMPLE = auto()
 400    TAG = auto()
 401    TEMPORARY = auto()
 402    TOP = auto()
 403    THEN = auto()
 404    TRUE = auto()
 405    TRUNCATE = auto()
 406    UNCACHE = auto()
 407    UNION = auto()
 408    UNNEST = auto()
 409    UNPIVOT = auto()
 410    UPDATE = auto()
 411    USE = auto()
 412    USING = auto()
 413    VALUES = auto()
 414    VIEW = auto()
 415    VOLATILE = auto()
 416    WHEN = auto()
 417    WHERE = auto()
 418    WINDOW = auto()
 419    WITH = auto()
 420    UNIQUE = auto()
 421    VERSION_SNAPSHOT = auto()
 422    TIMESTAMP_SNAPSHOT = auto()
 423    OPTION = auto()
 424    SINK = auto()
 425    SOURCE = auto()
 426    ANALYZE = auto()
 427    NAMESPACE = auto()
 428    EXPORT = auto()
 429
 430
 431_ALL_TOKEN_TYPES = list(TokenType)
 432_TOKEN_TYPE_TO_INDEX = {token_type: i for i, token_type in enumerate(_ALL_TOKEN_TYPES)}
 433
 434
 435class Token:
 436    __slots__ = ("token_type", "text", "line", "col", "start", "end", "comments")
 437
 438    @classmethod
 439    def number(cls, number: int) -> Token:
 440        """Returns a NUMBER token with `number` as its text."""
 441        return cls(TokenType.NUMBER, str(number))
 442
 443    @classmethod
 444    def string(cls, string: str) -> Token:
 445        """Returns a STRING token with `string` as its text."""
 446        return cls(TokenType.STRING, string)
 447
 448    @classmethod
 449    def identifier(cls, identifier: str) -> Token:
 450        """Returns an IDENTIFIER token with `identifier` as its text."""
 451        return cls(TokenType.IDENTIFIER, identifier)
 452
 453    @classmethod
 454    def var(cls, var: str) -> Token:
 455        """Returns an VAR token with `var` as its text."""
 456        return cls(TokenType.VAR, var)
 457
 458    def __init__(
 459        self,
 460        token_type: TokenType,
 461        text: str,
 462        line: int = 1,
 463        col: int = 1,
 464        start: int = 0,
 465        end: int = 0,
 466        comments: t.Optional[t.List[str]] = None,
 467    ) -> None:
 468        """Token initializer.
 469
 470        Args:
 471            token_type: The TokenType Enum.
 472            text: The text of the token.
 473            line: The line that the token ends on.
 474            col: The column that the token ends on.
 475            start: The start index of the token.
 476            end: The ending index of the token.
 477            comments: The comments to attach to the token.
 478        """
 479        self.token_type = token_type
 480        self.text = text
 481        self.line = line
 482        self.col = col
 483        self.start = start
 484        self.end = end
 485        self.comments = [] if comments is None else comments
 486
 487    def __repr__(self) -> str:
 488        attributes = ", ".join(f"{k}: {getattr(self, k)}" for k in self.__slots__)
 489        return f"<Token {attributes}>"
 490
 491
 492class _Tokenizer(type):
 493    def __new__(cls, clsname, bases, attrs):
 494        klass = super().__new__(cls, clsname, bases, attrs)
 495
 496        def _convert_quotes(arr: t.List[str | t.Tuple[str, str]]) -> t.Dict[str, str]:
 497            return dict(
 498                (item, item) if isinstance(item, str) else (item[0], item[1]) for item in arr
 499            )
 500
 501        def _quotes_to_format(
 502            token_type: TokenType, arr: t.List[str | t.Tuple[str, str]]
 503        ) -> t.Dict[str, t.Tuple[str, TokenType]]:
 504            return {k: (v, token_type) for k, v in _convert_quotes(arr).items()}
 505
 506        klass._QUOTES = _convert_quotes(klass.QUOTES)
 507        klass._IDENTIFIERS = _convert_quotes(klass.IDENTIFIERS)
 508
 509        klass._FORMAT_STRINGS = {
 510            **{
 511                p + s: (e, TokenType.NATIONAL_STRING)
 512                for s, e in klass._QUOTES.items()
 513                for p in ("n", "N")
 514            },
 515            **_quotes_to_format(TokenType.BIT_STRING, klass.BIT_STRINGS),
 516            **_quotes_to_format(TokenType.BYTE_STRING, klass.BYTE_STRINGS),
 517            **_quotes_to_format(TokenType.HEX_STRING, klass.HEX_STRINGS),
 518            **_quotes_to_format(TokenType.RAW_STRING, klass.RAW_STRINGS),
 519            **_quotes_to_format(TokenType.HEREDOC_STRING, klass.HEREDOC_STRINGS),
 520            **_quotes_to_format(TokenType.UNICODE_STRING, klass.UNICODE_STRINGS),
 521        }
 522
 523        klass._STRING_ESCAPES = set(klass.STRING_ESCAPES)
 524        klass._IDENTIFIER_ESCAPES = set(klass.IDENTIFIER_ESCAPES)
 525        klass._COMMENTS = {
 526            **dict(
 527                (comment, None) if isinstance(comment, str) else (comment[0], comment[1])
 528                for comment in klass.COMMENTS
 529            ),
 530            "{#": "#}",  # Ensure Jinja comments are tokenized correctly in all dialects
 531        }
 532        if klass.HINT_START in klass.KEYWORDS:
 533            klass._COMMENTS[klass.HINT_START] = "*/"
 534
 535        klass._KEYWORD_TRIE = new_trie(
 536            key.upper()
 537            for key in (
 538                *klass.KEYWORDS,
 539                *klass._COMMENTS,
 540                *klass._QUOTES,
 541                *klass._FORMAT_STRINGS,
 542            )
 543            if " " in key or any(single in key for single in klass.SINGLE_TOKENS)
 544        )
 545
 546        if USE_RS_TOKENIZER:
 547            settings = RsTokenizerSettings(
 548                white_space={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.WHITE_SPACE.items()},
 549                single_tokens={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.SINGLE_TOKENS.items()},
 550                keywords={k: _TOKEN_TYPE_TO_INDEX[v] for k, v in klass.KEYWORDS.items()},
 551                numeric_literals=klass.NUMERIC_LITERALS,
 552                identifiers=klass._IDENTIFIERS,
 553                identifier_escapes=klass._IDENTIFIER_ESCAPES,
 554                string_escapes=klass._STRING_ESCAPES,
 555                quotes=klass._QUOTES,
 556                format_strings={
 557                    k: (v1, _TOKEN_TYPE_TO_INDEX[v2])
 558                    for k, (v1, v2) in klass._FORMAT_STRINGS.items()
 559                },
 560                has_bit_strings=bool(klass.BIT_STRINGS),
 561                has_hex_strings=bool(klass.HEX_STRINGS),
 562                comments=klass._COMMENTS,
 563                var_single_tokens=klass.VAR_SINGLE_TOKENS,
 564                commands={_TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMANDS},
 565                command_prefix_tokens={
 566                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.COMMAND_PREFIX_TOKENS
 567                },
 568                heredoc_tag_is_identifier=klass.HEREDOC_TAG_IS_IDENTIFIER,
 569                string_escapes_allowed_in_raw_strings=klass.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS,
 570                nested_comments=klass.NESTED_COMMENTS,
 571                hint_start=klass.HINT_START,
 572                tokens_preceding_hint={
 573                    _TOKEN_TYPE_TO_INDEX[v] for v in klass.TOKENS_PRECEDING_HINT
 574                },
 575            )
 576            token_types = RsTokenTypeSettings(
 577                bit_string=_TOKEN_TYPE_TO_INDEX[TokenType.BIT_STRING],
 578                break_=_TOKEN_TYPE_TO_INDEX[TokenType.BREAK],
 579                dcolon=_TOKEN_TYPE_TO_INDEX[TokenType.DCOLON],
 580                heredoc_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEREDOC_STRING],
 581                raw_string=_TOKEN_TYPE_TO_INDEX[TokenType.RAW_STRING],
 582                hex_string=_TOKEN_TYPE_TO_INDEX[TokenType.HEX_STRING],
 583                identifier=_TOKEN_TYPE_TO_INDEX[TokenType.IDENTIFIER],
 584                number=_TOKEN_TYPE_TO_INDEX[TokenType.NUMBER],
 585                parameter=_TOKEN_TYPE_TO_INDEX[TokenType.PARAMETER],
 586                semicolon=_TOKEN_TYPE_TO_INDEX[TokenType.SEMICOLON],
 587                string=_TOKEN_TYPE_TO_INDEX[TokenType.STRING],
 588                var=_TOKEN_TYPE_TO_INDEX[TokenType.VAR],
 589                heredoc_string_alternative=_TOKEN_TYPE_TO_INDEX[klass.HEREDOC_STRING_ALTERNATIVE],
 590                hint=_TOKEN_TYPE_TO_INDEX[TokenType.HINT],
 591            )
 592            klass._RS_TOKENIZER = RsTokenizer(settings, token_types)
 593        else:
 594            klass._RS_TOKENIZER = None
 595
 596        return klass
 597
 598
 599class Tokenizer(metaclass=_Tokenizer):
 600    SINGLE_TOKENS = {
 601        "(": TokenType.L_PAREN,
 602        ")": TokenType.R_PAREN,
 603        "[": TokenType.L_BRACKET,
 604        "]": TokenType.R_BRACKET,
 605        "{": TokenType.L_BRACE,
 606        "}": TokenType.R_BRACE,
 607        "&": TokenType.AMP,
 608        "^": TokenType.CARET,
 609        ":": TokenType.COLON,
 610        ",": TokenType.COMMA,
 611        ".": TokenType.DOT,
 612        "-": TokenType.DASH,
 613        "=": TokenType.EQ,
 614        ">": TokenType.GT,
 615        "<": TokenType.LT,
 616        "%": TokenType.MOD,
 617        "!": TokenType.NOT,
 618        "|": TokenType.PIPE,
 619        "+": TokenType.PLUS,
 620        ";": TokenType.SEMICOLON,
 621        "/": TokenType.SLASH,
 622        "\\": TokenType.BACKSLASH,
 623        "*": TokenType.STAR,
 624        "~": TokenType.TILDA,
 625        "?": TokenType.PLACEHOLDER,
 626        "@": TokenType.PARAMETER,
 627        "#": TokenType.HASH,
 628        # Used for breaking a var like x'y' but nothing else the token type doesn't matter
 629        "'": TokenType.UNKNOWN,
 630        "`": TokenType.UNKNOWN,
 631        '"': TokenType.UNKNOWN,
 632    }
 633
 634    BIT_STRINGS: t.List[str | t.Tuple[str, str]] = []
 635    BYTE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 636    HEX_STRINGS: t.List[str | t.Tuple[str, str]] = []
 637    RAW_STRINGS: t.List[str | t.Tuple[str, str]] = []
 638    HEREDOC_STRINGS: t.List[str | t.Tuple[str, str]] = []
 639    UNICODE_STRINGS: t.List[str | t.Tuple[str, str]] = []
 640    IDENTIFIERS: t.List[str | t.Tuple[str, str]] = ['"']
 641    QUOTES: t.List[t.Tuple[str, str] | str] = ["'"]
 642    STRING_ESCAPES = ["'"]
 643    VAR_SINGLE_TOKENS: t.Set[str] = set()
 644
 645    # The strings in this list can always be used as escapes, regardless of the surrounding
 646    # identifier delimiters. By default, the closing delimiter is assumed to also act as an
 647    # identifier escape, e.g. if we use double-quotes, then they also act as escapes: "x"""
 648    IDENTIFIER_ESCAPES: t.List[str] = []
 649
 650    # Whether the heredoc tags follow the same lexical rules as unquoted identifiers
 651    HEREDOC_TAG_IS_IDENTIFIER = False
 652
 653    # Token that we'll generate as a fallback if the heredoc prefix doesn't correspond to a heredoc
 654    HEREDOC_STRING_ALTERNATIVE = TokenType.VAR
 655
 656    # Whether string escape characters function as such when placed within raw strings
 657    STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS = True
 658
 659    NESTED_COMMENTS = True
 660
 661    HINT_START = "/*+"
 662
 663    TOKENS_PRECEDING_HINT = {TokenType.SELECT, TokenType.INSERT, TokenType.UPDATE, TokenType.DELETE}
 664
 665    # Autofilled
 666    _COMMENTS: t.Dict[str, str] = {}
 667    _FORMAT_STRINGS: t.Dict[str, t.Tuple[str, TokenType]] = {}
 668    _IDENTIFIERS: t.Dict[str, str] = {}
 669    _IDENTIFIER_ESCAPES: t.Set[str] = set()
 670    _QUOTES: t.Dict[str, str] = {}
 671    _STRING_ESCAPES: t.Set[str] = set()
 672    _KEYWORD_TRIE: t.Dict = {}
 673    _RS_TOKENIZER: t.Optional[t.Any] = None
 674
 675    KEYWORDS: t.Dict[str, TokenType] = {
 676        **{f"{{%{postfix}": TokenType.BLOCK_START for postfix in ("", "+", "-")},
 677        **{f"{prefix}%}}": TokenType.BLOCK_END for prefix in ("", "+", "-")},
 678        **{f"{{{{{postfix}": TokenType.BLOCK_START for postfix in ("+", "-")},
 679        **{f"{prefix}}}}}": TokenType.BLOCK_END for prefix in ("+", "-")},
 680        HINT_START: TokenType.HINT,
 681        "==": TokenType.EQ,
 682        "::": TokenType.DCOLON,
 683        "||": TokenType.DPIPE,
 684        "|>": TokenType.PIPE_GT,
 685        ">=": TokenType.GTE,
 686        "<=": TokenType.LTE,
 687        "<>": TokenType.NEQ,
 688        "!=": TokenType.NEQ,
 689        ":=": TokenType.COLON_EQ,
 690        "<=>": TokenType.NULLSAFE_EQ,
 691        "->": TokenType.ARROW,
 692        "->>": TokenType.DARROW,
 693        "=>": TokenType.FARROW,
 694        "#>": TokenType.HASH_ARROW,
 695        "#>>": TokenType.DHASH_ARROW,
 696        "<->": TokenType.LR_ARROW,
 697        "&&": TokenType.DAMP,
 698        "??": TokenType.DQMARK,
 699        "~~~": TokenType.GLOB,
 700        "~~": TokenType.LIKE,
 701        "~~*": TokenType.ILIKE,
 702        "~*": TokenType.IRLIKE,
 703        "ALL": TokenType.ALL,
 704        "ALWAYS": TokenType.ALWAYS,
 705        "AND": TokenType.AND,
 706        "ANTI": TokenType.ANTI,
 707        "ANY": TokenType.ANY,
 708        "ASC": TokenType.ASC,
 709        "AS": TokenType.ALIAS,
 710        "ASOF": TokenType.ASOF,
 711        "AUTOINCREMENT": TokenType.AUTO_INCREMENT,
 712        "AUTO_INCREMENT": TokenType.AUTO_INCREMENT,
 713        "BEGIN": TokenType.BEGIN,
 714        "BETWEEN": TokenType.BETWEEN,
 715        "CACHE": TokenType.CACHE,
 716        "UNCACHE": TokenType.UNCACHE,
 717        "CASE": TokenType.CASE,
 718        "CHARACTER SET": TokenType.CHARACTER_SET,
 719        "CLUSTER BY": TokenType.CLUSTER_BY,
 720        "COLLATE": TokenType.COLLATE,
 721        "COLUMN": TokenType.COLUMN,
 722        "COMMIT": TokenType.COMMIT,
 723        "CONNECT BY": TokenType.CONNECT_BY,
 724        "CONSTRAINT": TokenType.CONSTRAINT,
 725        "COPY": TokenType.COPY,
 726        "CREATE": TokenType.CREATE,
 727        "CROSS": TokenType.CROSS,
 728        "CUBE": TokenType.CUBE,
 729        "CURRENT_DATE": TokenType.CURRENT_DATE,
 730        "CURRENT_SCHEMA": TokenType.CURRENT_SCHEMA,
 731        "CURRENT_TIME": TokenType.CURRENT_TIME,
 732        "CURRENT_TIMESTAMP": TokenType.CURRENT_TIMESTAMP,
 733        "CURRENT_USER": TokenType.CURRENT_USER,
 734        "DATABASE": TokenType.DATABASE,
 735        "DEFAULT": TokenType.DEFAULT,
 736        "DELETE": TokenType.DELETE,
 737        "DESC": TokenType.DESC,
 738        "DESCRIBE": TokenType.DESCRIBE,
 739        "DISTINCT": TokenType.DISTINCT,
 740        "DISTRIBUTE BY": TokenType.DISTRIBUTE_BY,
 741        "DIV": TokenType.DIV,
 742        "DROP": TokenType.DROP,
 743        "ELSE": TokenType.ELSE,
 744        "END": TokenType.END,
 745        "ENUM": TokenType.ENUM,
 746        "ESCAPE": TokenType.ESCAPE,
 747        "EXCEPT": TokenType.EXCEPT,
 748        "EXECUTE": TokenType.EXECUTE,
 749        "EXISTS": TokenType.EXISTS,
 750        "FALSE": TokenType.FALSE,
 751        "FETCH": TokenType.FETCH,
 752        "FILTER": TokenType.FILTER,
 753        "FIRST": TokenType.FIRST,
 754        "FULL": TokenType.FULL,
 755        "FUNCTION": TokenType.FUNCTION,
 756        "FOR": TokenType.FOR,
 757        "FOREIGN KEY": TokenType.FOREIGN_KEY,
 758        "FORMAT": TokenType.FORMAT,
 759        "FROM": TokenType.FROM,
 760        "GEOGRAPHY": TokenType.GEOGRAPHY,
 761        "GEOMETRY": TokenType.GEOMETRY,
 762        "GLOB": TokenType.GLOB,
 763        "GROUP BY": TokenType.GROUP_BY,
 764        "GROUPING SETS": TokenType.GROUPING_SETS,
 765        "HAVING": TokenType.HAVING,
 766        "ILIKE": TokenType.ILIKE,
 767        "IN": TokenType.IN,
 768        "INDEX": TokenType.INDEX,
 769        "INET": TokenType.INET,
 770        "INNER": TokenType.INNER,
 771        "INSERT": TokenType.INSERT,
 772        "INTERVAL": TokenType.INTERVAL,
 773        "INTERSECT": TokenType.INTERSECT,
 774        "INTO": TokenType.INTO,
 775        "IS": TokenType.IS,
 776        "ISNULL": TokenType.ISNULL,
 777        "JOIN": TokenType.JOIN,
 778        "KEEP": TokenType.KEEP,
 779        "KILL": TokenType.KILL,
 780        "LATERAL": TokenType.LATERAL,
 781        "LEFT": TokenType.LEFT,
 782        "LIKE": TokenType.LIKE,
 783        "LIMIT": TokenType.LIMIT,
 784        "LOAD": TokenType.LOAD,
 785        "LOCK": TokenType.LOCK,
 786        "MERGE": TokenType.MERGE,
 787        "NAMESPACE": TokenType.NAMESPACE,
 788        "NATURAL": TokenType.NATURAL,
 789        "NEXT": TokenType.NEXT,
 790        "NOT": TokenType.NOT,
 791        "NOTNULL": TokenType.NOTNULL,
 792        "NULL": TokenType.NULL,
 793        "OBJECT": TokenType.OBJECT,
 794        "OFFSET": TokenType.OFFSET,
 795        "ON": TokenType.ON,
 796        "OR": TokenType.OR,
 797        "XOR": TokenType.XOR,
 798        "ORDER BY": TokenType.ORDER_BY,
 799        "ORDINALITY": TokenType.ORDINALITY,
 800        "OUTER": TokenType.OUTER,
 801        "OVER": TokenType.OVER,
 802        "OVERLAPS": TokenType.OVERLAPS,
 803        "OVERWRITE": TokenType.OVERWRITE,
 804        "PARTITION": TokenType.PARTITION,
 805        "PARTITION BY": TokenType.PARTITION_BY,
 806        "PARTITIONED BY": TokenType.PARTITION_BY,
 807        "PARTITIONED_BY": TokenType.PARTITION_BY,
 808        "PERCENT": TokenType.PERCENT,
 809        "PIVOT": TokenType.PIVOT,
 810        "PRAGMA": TokenType.PRAGMA,
 811        "PRIMARY KEY": TokenType.PRIMARY_KEY,
 812        "PROCEDURE": TokenType.PROCEDURE,
 813        "QUALIFY": TokenType.QUALIFY,
 814        "RANGE": TokenType.RANGE,
 815        "RECURSIVE": TokenType.RECURSIVE,
 816        "REGEXP": TokenType.RLIKE,
 817        "RENAME": TokenType.RENAME,
 818        "REPLACE": TokenType.REPLACE,
 819        "RETURNING": TokenType.RETURNING,
 820        "REFERENCES": TokenType.REFERENCES,
 821        "RIGHT": TokenType.RIGHT,
 822        "RLIKE": TokenType.RLIKE,
 823        "ROLLBACK": TokenType.ROLLBACK,
 824        "ROLLUP": TokenType.ROLLUP,
 825        "ROW": TokenType.ROW,
 826        "ROWS": TokenType.ROWS,
 827        "SCHEMA": TokenType.SCHEMA,
 828        "SELECT": TokenType.SELECT,
 829        "SEMI": TokenType.SEMI,
 830        "SET": TokenType.SET,
 831        "SETTINGS": TokenType.SETTINGS,
 832        "SHOW": TokenType.SHOW,
 833        "SIMILAR TO": TokenType.SIMILAR_TO,
 834        "SOME": TokenType.SOME,
 835        "SORT BY": TokenType.SORT_BY,
 836        "START WITH": TokenType.START_WITH,
 837        "STRAIGHT_JOIN": TokenType.STRAIGHT_JOIN,
 838        "TABLE": TokenType.TABLE,
 839        "TABLESAMPLE": TokenType.TABLE_SAMPLE,
 840        "TEMP": TokenType.TEMPORARY,
 841        "TEMPORARY": TokenType.TEMPORARY,
 842        "THEN": TokenType.THEN,
 843        "TRUE": TokenType.TRUE,
 844        "TRUNCATE": TokenType.TRUNCATE,
 845        "UNION": TokenType.UNION,
 846        "UNKNOWN": TokenType.UNKNOWN,
 847        "UNNEST": TokenType.UNNEST,
 848        "UNPIVOT": TokenType.UNPIVOT,
 849        "UPDATE": TokenType.UPDATE,
 850        "USE": TokenType.USE,
 851        "USING": TokenType.USING,
 852        "UUID": TokenType.UUID,
 853        "VALUES": TokenType.VALUES,
 854        "VIEW": TokenType.VIEW,
 855        "VOLATILE": TokenType.VOLATILE,
 856        "WHEN": TokenType.WHEN,
 857        "WHERE": TokenType.WHERE,
 858        "WINDOW": TokenType.WINDOW,
 859        "WITH": TokenType.WITH,
 860        "APPLY": TokenType.APPLY,
 861        "ARRAY": TokenType.ARRAY,
 862        "BIT": TokenType.BIT,
 863        "BOOL": TokenType.BOOLEAN,
 864        "BOOLEAN": TokenType.BOOLEAN,
 865        "BYTE": TokenType.TINYINT,
 866        "MEDIUMINT": TokenType.MEDIUMINT,
 867        "INT1": TokenType.TINYINT,
 868        "TINYINT": TokenType.TINYINT,
 869        "INT16": TokenType.SMALLINT,
 870        "SHORT": TokenType.SMALLINT,
 871        "SMALLINT": TokenType.SMALLINT,
 872        "HUGEINT": TokenType.INT128,
 873        "UHUGEINT": TokenType.UINT128,
 874        "INT2": TokenType.SMALLINT,
 875        "INTEGER": TokenType.INT,
 876        "INT": TokenType.INT,
 877        "INT4": TokenType.INT,
 878        "INT32": TokenType.INT,
 879        "INT64": TokenType.BIGINT,
 880        "INT128": TokenType.INT128,
 881        "INT256": TokenType.INT256,
 882        "LONG": TokenType.BIGINT,
 883        "BIGINT": TokenType.BIGINT,
 884        "INT8": TokenType.TINYINT,
 885        "UINT": TokenType.UINT,
 886        "UINT128": TokenType.UINT128,
 887        "UINT256": TokenType.UINT256,
 888        "DEC": TokenType.DECIMAL,
 889        "DECIMAL": TokenType.DECIMAL,
 890        "DECIMAL32": TokenType.DECIMAL32,
 891        "DECIMAL64": TokenType.DECIMAL64,
 892        "DECIMAL128": TokenType.DECIMAL128,
 893        "DECIMAL256": TokenType.DECIMAL256,
 894        "BIGDECIMAL": TokenType.BIGDECIMAL,
 895        "BIGNUMERIC": TokenType.BIGDECIMAL,
 896        "LIST": TokenType.LIST,
 897        "MAP": TokenType.MAP,
 898        "NULLABLE": TokenType.NULLABLE,
 899        "NUMBER": TokenType.DECIMAL,
 900        "NUMERIC": TokenType.DECIMAL,
 901        "FIXED": TokenType.DECIMAL,
 902        "REAL": TokenType.FLOAT,
 903        "FLOAT": TokenType.FLOAT,
 904        "FLOAT4": TokenType.FLOAT,
 905        "FLOAT8": TokenType.DOUBLE,
 906        "DOUBLE": TokenType.DOUBLE,
 907        "DOUBLE PRECISION": TokenType.DOUBLE,
 908        "JSON": TokenType.JSON,
 909        "JSONB": TokenType.JSONB,
 910        "CHAR": TokenType.CHAR,
 911        "CHARACTER": TokenType.CHAR,
 912        "CHAR VARYING": TokenType.VARCHAR,
 913        "CHARACTER VARYING": TokenType.VARCHAR,
 914        "NCHAR": TokenType.NCHAR,
 915        "VARCHAR": TokenType.VARCHAR,
 916        "VARCHAR2": TokenType.VARCHAR,
 917        "NVARCHAR": TokenType.NVARCHAR,
 918        "NVARCHAR2": TokenType.NVARCHAR,
 919        "BPCHAR": TokenType.BPCHAR,
 920        "STR": TokenType.TEXT,
 921        "STRING": TokenType.TEXT,
 922        "TEXT": TokenType.TEXT,
 923        "LONGTEXT": TokenType.LONGTEXT,
 924        "MEDIUMTEXT": TokenType.MEDIUMTEXT,
 925        "TINYTEXT": TokenType.TINYTEXT,
 926        "CLOB": TokenType.TEXT,
 927        "LONGVARCHAR": TokenType.TEXT,
 928        "BINARY": TokenType.BINARY,
 929        "BLOB": TokenType.VARBINARY,
 930        "LONGBLOB": TokenType.LONGBLOB,
 931        "MEDIUMBLOB": TokenType.MEDIUMBLOB,
 932        "TINYBLOB": TokenType.TINYBLOB,
 933        "BYTEA": TokenType.VARBINARY,
 934        "VARBINARY": TokenType.VARBINARY,
 935        "TIME": TokenType.TIME,
 936        "TIMETZ": TokenType.TIMETZ,
 937        "TIMESTAMP": TokenType.TIMESTAMP,
 938        "TIMESTAMPTZ": TokenType.TIMESTAMPTZ,
 939        "TIMESTAMPLTZ": TokenType.TIMESTAMPLTZ,
 940        "TIMESTAMP_LTZ": TokenType.TIMESTAMPLTZ,
 941        "TIMESTAMPNTZ": TokenType.TIMESTAMPNTZ,
 942        "TIMESTAMP_NTZ": TokenType.TIMESTAMPNTZ,
 943        "DATE": TokenType.DATE,
 944        "DATETIME": TokenType.DATETIME,
 945        "INT4RANGE": TokenType.INT4RANGE,
 946        "INT4MULTIRANGE": TokenType.INT4MULTIRANGE,
 947        "INT8RANGE": TokenType.INT8RANGE,
 948        "INT8MULTIRANGE": TokenType.INT8MULTIRANGE,
 949        "NUMRANGE": TokenType.NUMRANGE,
 950        "NUMMULTIRANGE": TokenType.NUMMULTIRANGE,
 951        "TSRANGE": TokenType.TSRANGE,
 952        "TSMULTIRANGE": TokenType.TSMULTIRANGE,
 953        "TSTZRANGE": TokenType.TSTZRANGE,
 954        "TSTZMULTIRANGE": TokenType.TSTZMULTIRANGE,
 955        "DATERANGE": TokenType.DATERANGE,
 956        "DATEMULTIRANGE": TokenType.DATEMULTIRANGE,
 957        "UNIQUE": TokenType.UNIQUE,
 958        "VECTOR": TokenType.VECTOR,
 959        "STRUCT": TokenType.STRUCT,
 960        "SEQUENCE": TokenType.SEQUENCE,
 961        "VARIANT": TokenType.VARIANT,
 962        "ALTER": TokenType.ALTER,
 963        "ANALYZE": TokenType.ANALYZE,
 964        "CALL": TokenType.COMMAND,
 965        "COMMENT": TokenType.COMMENT,
 966        "EXPLAIN": TokenType.COMMAND,
 967        "GRANT": TokenType.GRANT,
 968        "OPTIMIZE": TokenType.COMMAND,
 969        "PREPARE": TokenType.COMMAND,
 970        "VACUUM": TokenType.COMMAND,
 971        "USER-DEFINED": TokenType.USERDEFINED,
 972        "FOR VERSION": TokenType.VERSION_SNAPSHOT,
 973        "FOR TIMESTAMP": TokenType.TIMESTAMP_SNAPSHOT,
 974    }
 975
 976    WHITE_SPACE: t.Dict[t.Optional[str], TokenType] = {
 977        " ": TokenType.SPACE,
 978        "\t": TokenType.SPACE,
 979        "\n": TokenType.BREAK,
 980        "\r": TokenType.BREAK,
 981    }
 982
 983    COMMANDS = {
 984        TokenType.COMMAND,
 985        TokenType.EXECUTE,
 986        TokenType.FETCH,
 987        TokenType.SHOW,
 988        TokenType.RENAME,
 989    }
 990
 991    COMMAND_PREFIX_TOKENS = {TokenType.SEMICOLON, TokenType.BEGIN}
 992
 993    # Handle numeric literals like in hive (3L = BIGINT)
 994    NUMERIC_LITERALS: t.Dict[str, str] = {}
 995
 996    COMMENTS = ["--", ("/*", "*/")]
 997
 998    __slots__ = (
 999        "sql",
1000        "size",
1001        "tokens",
1002        "dialect",
1003        "use_rs_tokenizer",
1004        "_start",
1005        "_current",
1006        "_line",
1007        "_col",
1008        "_comments",
1009        "_char",
1010        "_end",
1011        "_peek",
1012        "_prev_token_line",
1013        "_rs_dialect_settings",
1014    )
1015
1016    def __init__(
1017        self, dialect: DialectType = None, use_rs_tokenizer: t.Optional[bool] = None
1018    ) -> None:
1019        from sqlglot.dialects import Dialect
1020
1021        self.dialect = Dialect.get_or_raise(dialect)
1022
1023        # initialize `use_rs_tokenizer`, and allow it to be overwritten per Tokenizer instance
1024        self.use_rs_tokenizer = (
1025            use_rs_tokenizer if use_rs_tokenizer is not None else USE_RS_TOKENIZER
1026        )
1027
1028        if self.use_rs_tokenizer:
1029            self._rs_dialect_settings = RsTokenizerDialectSettings(
1030                unescaped_sequences=self.dialect.UNESCAPED_SEQUENCES,
1031                identifiers_can_start_with_digit=self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT,
1032                numbers_can_be_underscore_separated=self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED,
1033            )
1034
1035        self.reset()
1036
1037    def reset(self) -> None:
1038        self.sql = ""
1039        self.size = 0
1040        self.tokens: t.List[Token] = []
1041        self._start = 0
1042        self._current = 0
1043        self._line = 1
1044        self._col = 0
1045        self._comments: t.List[str] = []
1046
1047        self._char = ""
1048        self._end = False
1049        self._peek = ""
1050        self._prev_token_line = -1
1051
1052    def tokenize(self, sql: str) -> t.List[Token]:
1053        """Returns a list of tokens corresponding to the SQL string `sql`."""
1054        if self.use_rs_tokenizer:
1055            return self.tokenize_rs(sql)
1056
1057        self.reset()
1058        self.sql = sql
1059        self.size = len(sql)
1060
1061        try:
1062            self._scan()
1063        except Exception as e:
1064            start = max(self._current - 50, 0)
1065            end = min(self._current + 50, self.size - 1)
1066            context = self.sql[start:end]
1067            raise TokenError(f"Error tokenizing '{context}'") from e
1068
1069        return self.tokens
1070
1071    def _scan(self, until: t.Optional[t.Callable] = None) -> None:
1072        while self.size and not self._end:
1073            current = self._current
1074
1075            # Skip spaces here rather than iteratively calling advance() for performance reasons
1076            while current < self.size:
1077                char = self.sql[current]
1078
1079                if char.isspace() and (char == " " or char == "\t"):
1080                    current += 1
1081                else:
1082                    break
1083
1084            offset = current - self._current if current > self._current else 1
1085
1086            self._start = current
1087            self._advance(offset)
1088
1089            if not self._char.isspace():
1090                if self._char.isdigit():
1091                    self._scan_number()
1092                elif self._char in self._IDENTIFIERS:
1093                    self._scan_identifier(self._IDENTIFIERS[self._char])
1094                else:
1095                    self._scan_keywords()
1096
1097            if until and until():
1098                break
1099
1100        if self.tokens and self._comments:
1101            self.tokens[-1].comments.extend(self._comments)
1102
1103    def _chars(self, size: int) -> str:
1104        if size == 1:
1105            return self._char
1106
1107        start = self._current - 1
1108        end = start + size
1109
1110        return self.sql[start:end] if end <= self.size else ""
1111
1112    def _advance(self, i: int = 1, alnum: bool = False) -> None:
1113        if self.WHITE_SPACE.get(self._char) is TokenType.BREAK:
1114            # Ensures we don't count an extra line if we get a \r\n line break sequence
1115            if not (self._char == "\r" and self._peek == "\n"):
1116                self._col = i
1117                self._line += 1
1118        else:
1119            self._col += i
1120
1121        self._current += i
1122        self._end = self._current >= self.size
1123        self._char = self.sql[self._current - 1]
1124        self._peek = "" if self._end else self.sql[self._current]
1125
1126        if alnum and self._char.isalnum():
1127            # Here we use local variables instead of attributes for better performance
1128            _col = self._col
1129            _current = self._current
1130            _end = self._end
1131            _peek = self._peek
1132
1133            while _peek.isalnum():
1134                _col += 1
1135                _current += 1
1136                _end = _current >= self.size
1137                _peek = "" if _end else self.sql[_current]
1138
1139            self._col = _col
1140            self._current = _current
1141            self._end = _end
1142            self._peek = _peek
1143            self._char = self.sql[_current - 1]
1144
1145    @property
1146    def _text(self) -> str:
1147        return self.sql[self._start : self._current]
1148
1149    def _add(self, token_type: TokenType, text: t.Optional[str] = None) -> None:
1150        self._prev_token_line = self._line
1151
1152        if self._comments and token_type == TokenType.SEMICOLON and self.tokens:
1153            self.tokens[-1].comments.extend(self._comments)
1154            self._comments = []
1155
1156        self.tokens.append(
1157            Token(
1158                token_type,
1159                text=self._text if text is None else text,
1160                line=self._line,
1161                col=self._col,
1162                start=self._start,
1163                end=self._current - 1,
1164                comments=self._comments,
1165            )
1166        )
1167        self._comments = []
1168
1169        # If we have either a semicolon or a begin token before the command's token, we'll parse
1170        # whatever follows the command's token as a string
1171        if (
1172            token_type in self.COMMANDS
1173            and self._peek != ";"
1174            and (len(self.tokens) == 1 or self.tokens[-2].token_type in self.COMMAND_PREFIX_TOKENS)
1175        ):
1176            start = self._current
1177            tokens = len(self.tokens)
1178            self._scan(lambda: self._peek == ";")
1179            self.tokens = self.tokens[:tokens]
1180            text = self.sql[start : self._current].strip()
1181            if text:
1182                self._add(TokenType.STRING, text)
1183
1184    def _scan_keywords(self) -> None:
1185        size = 0
1186        word = None
1187        chars = self._text
1188        char = chars
1189        prev_space = False
1190        skip = False
1191        trie = self._KEYWORD_TRIE
1192        single_token = char in self.SINGLE_TOKENS
1193
1194        while chars:
1195            if skip:
1196                result = TrieResult.PREFIX
1197            else:
1198                result, trie = in_trie(trie, char.upper())
1199
1200            if result == TrieResult.FAILED:
1201                break
1202            if result == TrieResult.EXISTS:
1203                word = chars
1204
1205            end = self._current + size
1206            size += 1
1207
1208            if end < self.size:
1209                char = self.sql[end]
1210                single_token = single_token or char in self.SINGLE_TOKENS
1211                is_space = char.isspace()
1212
1213                if not is_space or not prev_space:
1214                    if is_space:
1215                        char = " "
1216                    chars += char
1217                    prev_space = is_space
1218                    skip = False
1219                else:
1220                    skip = True
1221            else:
1222                char = ""
1223                break
1224
1225        if word:
1226            if self._scan_string(word):
1227                return
1228            if self._scan_comment(word):
1229                return
1230            if prev_space or single_token or not char:
1231                self._advance(size - 1)
1232                word = word.upper()
1233                self._add(self.KEYWORDS[word], text=word)
1234                return
1235
1236        if self._char in self.SINGLE_TOKENS:
1237            self._add(self.SINGLE_TOKENS[self._char], text=self._char)
1238            return
1239
1240        self._scan_var()
1241
1242    def _scan_comment(self, comment_start: str) -> bool:
1243        if comment_start not in self._COMMENTS:
1244            return False
1245
1246        comment_start_line = self._line
1247        comment_start_size = len(comment_start)
1248        comment_end = self._COMMENTS[comment_start]
1249
1250        if comment_end:
1251            # Skip the comment's start delimiter
1252            self._advance(comment_start_size)
1253
1254            comment_count = 1
1255            comment_end_size = len(comment_end)
1256
1257            while not self._end:
1258                if self._chars(comment_end_size) == comment_end:
1259                    comment_count -= 1
1260                    if not comment_count:
1261                        break
1262
1263                self._advance(alnum=True)
1264
1265                # Nested comments are allowed by some dialects, e.g. databricks, duckdb, postgres
1266                if (
1267                    self.NESTED_COMMENTS
1268                    and not self._end
1269                    and self._chars(comment_end_size) == comment_start
1270                ):
1271                    self._advance(comment_start_size)
1272                    comment_count += 1
1273
1274            self._comments.append(self._text[comment_start_size : -comment_end_size + 1])
1275            self._advance(comment_end_size - 1)
1276        else:
1277            while not self._end and self.WHITE_SPACE.get(self._peek) is not TokenType.BREAK:
1278                self._advance(alnum=True)
1279            self._comments.append(self._text[comment_start_size:])
1280
1281        if (
1282            comment_start == self.HINT_START
1283            and self.tokens
1284            and self.tokens[-1].token_type in self.TOKENS_PRECEDING_HINT
1285        ):
1286            self._add(TokenType.HINT)
1287
1288        # Leading comment is attached to the succeeding token, whilst trailing comment to the preceding.
1289        # Multiple consecutive comments are preserved by appending them to the current comments list.
1290        if comment_start_line == self._prev_token_line:
1291            self.tokens[-1].comments.extend(self._comments)
1292            self._comments = []
1293            self._prev_token_line = self._line
1294
1295        return True
1296
1297    def _scan_number(self) -> None:
1298        if self._char == "0":
1299            peek = self._peek.upper()
1300            if peek == "B":
1301                return self._scan_bits() if self.BIT_STRINGS else self._add(TokenType.NUMBER)
1302            elif peek == "X":
1303                return self._scan_hex() if self.HEX_STRINGS else self._add(TokenType.NUMBER)
1304
1305        decimal = False
1306        scientific = 0
1307
1308        while True:
1309            if self._peek.isdigit():
1310                self._advance()
1311            elif self._peek == "." and not decimal:
1312                if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER:
1313                    return self._add(TokenType.NUMBER)
1314                decimal = True
1315                self._advance()
1316            elif self._peek in ("-", "+") and scientific == 1:
1317                scientific += 1
1318                self._advance()
1319            elif self._peek.upper() == "E" and not scientific:
1320                scientific += 1
1321                self._advance()
1322            elif self._peek.isidentifier():
1323                number_text = self._text
1324                literal = ""
1325
1326                while self._peek.strip() and self._peek not in self.SINGLE_TOKENS:
1327                    literal += self._peek
1328                    self._advance()
1329
1330                token_type = self.KEYWORDS.get(self.NUMERIC_LITERALS.get(literal.upper(), ""))
1331
1332                if token_type:
1333                    self._add(TokenType.NUMBER, number_text)
1334                    self._add(TokenType.DCOLON, "::")
1335                    return self._add(token_type, literal)
1336                else:
1337                    replaced = literal.replace("_", "")
1338                    if self.dialect.NUMBERS_CAN_BE_UNDERSCORE_SEPARATED and replaced.isdigit():
1339                        return self._add(TokenType.NUMBER, number_text + replaced)
1340                    if self.dialect.IDENTIFIERS_CAN_START_WITH_DIGIT:
1341                        return self._add(TokenType.VAR)
1342
1343                self._advance(-len(literal))
1344                return self._add(TokenType.NUMBER, number_text)
1345            else:
1346                return self._add(TokenType.NUMBER)
1347
1348    def _scan_bits(self) -> None:
1349        self._advance()
1350        value = self._extract_value()
1351        try:
1352            # If `value` can't be converted to a binary, fallback to tokenizing it as an identifier
1353            int(value, 2)
1354            self._add(TokenType.BIT_STRING, value[2:])  # Drop the 0b
1355        except ValueError:
1356            self._add(TokenType.IDENTIFIER)
1357
1358    def _scan_hex(self) -> None:
1359        self._advance()
1360        value = self._extract_value()
1361        try:
1362            # If `value` can't be converted to a hex, fallback to tokenizing it as an identifier
1363            int(value, 16)
1364            self._add(TokenType.HEX_STRING, value[2:])  # Drop the 0x
1365        except ValueError:
1366            self._add(TokenType.IDENTIFIER)
1367
1368    def _extract_value(self) -> str:
1369        while True:
1370            char = self._peek.strip()
1371            if char and char not in self.SINGLE_TOKENS:
1372                self._advance(alnum=True)
1373            else:
1374                break
1375
1376        return self._text
1377
1378    def _scan_string(self, start: str) -> bool:
1379        base = None
1380        token_type = TokenType.STRING
1381
1382        if start in self._QUOTES:
1383            end = self._QUOTES[start]
1384        elif start in self._FORMAT_STRINGS:
1385            end, token_type = self._FORMAT_STRINGS[start]
1386
1387            if token_type == TokenType.HEX_STRING:
1388                base = 16
1389            elif token_type == TokenType.BIT_STRING:
1390                base = 2
1391            elif token_type == TokenType.HEREDOC_STRING:
1392                self._advance()
1393
1394                if self._char == end:
1395                    tag = ""
1396                else:
1397                    tag = self._extract_string(
1398                        end,
1399                        raw_string=True,
1400                        raise_unmatched=not self.HEREDOC_TAG_IS_IDENTIFIER,
1401                    )
1402
1403                if tag and self.HEREDOC_TAG_IS_IDENTIFIER and (self._end or not tag.isidentifier()):
1404                    if not self._end:
1405                        self._advance(-1)
1406
1407                    self._advance(-len(tag))
1408                    self._add(self.HEREDOC_STRING_ALTERNATIVE)
1409                    return True
1410
1411                end = f"{start}{tag}{end}"
1412        else:
1413            return False
1414
1415        self._advance(len(start))
1416        text = self._extract_string(end, raw_string=token_type == TokenType.RAW_STRING)
1417
1418        if base:
1419            try:
1420                int(text, base)
1421            except Exception:
1422                raise TokenError(
1423                    f"Numeric string contains invalid characters from {self._line}:{self._start}"
1424                )
1425
1426        self._add(token_type, text)
1427        return True
1428
1429    def _scan_identifier(self, identifier_end: str) -> None:
1430        self._advance()
1431        text = self._extract_string(
1432            identifier_end, escapes=self._IDENTIFIER_ESCAPES | {identifier_end}
1433        )
1434        self._add(TokenType.IDENTIFIER, text)
1435
1436    def _scan_var(self) -> None:
1437        while True:
1438            char = self._peek.strip()
1439            if char and (char in self.VAR_SINGLE_TOKENS or char not in self.SINGLE_TOKENS):
1440                self._advance(alnum=True)
1441            else:
1442                break
1443
1444        self._add(
1445            TokenType.VAR
1446            if self.tokens and self.tokens[-1].token_type == TokenType.PARAMETER
1447            else self.KEYWORDS.get(self._text.upper(), TokenType.VAR)
1448        )
1449
1450    def _extract_string(
1451        self,
1452        delimiter: str,
1453        escapes: t.Optional[t.Set[str]] = None,
1454        raw_string: bool = False,
1455        raise_unmatched: bool = True,
1456    ) -> str:
1457        text = ""
1458        delim_size = len(delimiter)
1459        escapes = self._STRING_ESCAPES if escapes is None else escapes
1460
1461        while True:
1462            if (
1463                not raw_string
1464                and self.dialect.UNESCAPED_SEQUENCES
1465                and self._peek
1466                and self._char in self.STRING_ESCAPES
1467            ):
1468                unescaped_sequence = self.dialect.UNESCAPED_SEQUENCES.get(self._char + self._peek)
1469                if unescaped_sequence:
1470                    self._advance(2)
1471                    text += unescaped_sequence
1472                    continue
1473            if (
1474                (self.STRING_ESCAPES_ALLOWED_IN_RAW_STRINGS or not raw_string)
1475                and self._char in escapes
1476                and (self._peek == delimiter or self._peek in escapes)
1477                and (self._char not in self._QUOTES or self._char == self._peek)
1478            ):
1479                if self._peek == delimiter:
1480                    text += self._peek
1481                else:
1482                    text += self._char + self._peek
1483
1484                if self._current + 1 < self.size:
1485                    self._advance(2)
1486                else:
1487                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._current}")
1488            else:
1489                if self._chars(delim_size) == delimiter:
1490                    if delim_size > 1:
1491                        self._advance(delim_size - 1)
1492                    break
1493
1494                if self._end:
1495                    if not raise_unmatched:
1496                        return text + self._char
1497
1498                    raise TokenError(f"Missing {delimiter} from {self._line}:{self._start}")
1499
1500                current = self._current - 1
1501                self._advance(alnum=True)
1502                text += self.sql[current : self._current - 1]
1503
1504        return text
1505
1506    def tokenize_rs(self, sql: str) -> t.List[Token]:
1507        if not self._RS_TOKENIZER:
1508            raise SqlglotError("Rust tokenizer is not available")
1509
1510        tokens, error_msg = self._RS_TOKENIZER.tokenize(sql, self._rs_dialect_settings)
1511        for token in tokens:
1512            token.token_type = _ALL_TOKEN_TYPES[token.token_type_index]
1513
1514        # Setting this here so partial token lists can be inspected even if there is a failure
1515        self.tokens = tokens
1516
1517        if error_msg is not None:
1518            raise TokenError(error_msg)
1519
1520        return tokens