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