sqlglot.tokens

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