sqlglot.tokens

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