diff options
Diffstat (limited to 'sqlglot/parser.py')
| -rw-r--r-- | sqlglot/parser.py | 157 |
1 files changed, 122 insertions, 35 deletions
diff --git a/sqlglot/parser.py b/sqlglot/parser.py index 90fdade..a36251e 100644 --- a/sqlglot/parser.py +++ b/sqlglot/parser.py @@ -96,6 +96,7 @@ class Parser(metaclass=_Parser): NO_PAREN_FUNCTIONS = { TokenType.CURRENT_DATE: exp.CurrentDate, TokenType.CURRENT_DATETIME: exp.CurrentDate, + TokenType.CURRENT_TIME: exp.CurrentTime, TokenType.CURRENT_TIMESTAMP: exp.CurrentTimestamp, } @@ -198,7 +199,6 @@ class Parser(metaclass=_Parser): TokenType.COMMIT, TokenType.COMPOUND, TokenType.CONSTRAINT, - TokenType.CURRENT_TIME, TokenType.DEFAULT, TokenType.DELETE, TokenType.DESCRIBE, @@ -370,8 +370,9 @@ class Parser(metaclass=_Parser): LAMBDAS = { TokenType.ARROW: lambda self, expressions: self.expression( exp.Lambda, - this=self._parse_conjunction().transform( - self._replace_lambda, {node.name for node in expressions} + this=self._replace_lambda( + self._parse_conjunction(), + {node.name for node in expressions}, ), expressions=expressions, ), @@ -441,6 +442,7 @@ class Parser(metaclass=_Parser): exp.With: lambda self: self._parse_with(), exp.Window: lambda self: self._parse_named_window(), exp.Qualify: lambda self: self._parse_qualify(), + exp.Returning: lambda self: self._parse_returning(), "JOIN_TYPE": lambda self: self._parse_join_side_and_kind(), } @@ -460,6 +462,7 @@ class Parser(metaclass=_Parser): TokenType.LOAD_DATA: lambda self: self._parse_load_data(), TokenType.MERGE: lambda self: self._parse_merge(), TokenType.ROLLBACK: lambda self: self._parse_commit_or_rollback(), + TokenType.SET: lambda self: self._parse_set(), TokenType.UNCACHE: lambda self: self._parse_uncache(), TokenType.UPDATE: lambda self: self._parse_update(), TokenType.USE: lambda self: self.expression( @@ -656,15 +659,15 @@ class Parser(metaclass=_Parser): } FUNCTION_PARSERS: t.Dict[str, t.Callable] = { + "CAST": lambda self: self._parse_cast(self.STRICT_CAST), "CONVERT": lambda self: self._parse_convert(self.STRICT_CAST), - "TRY_CONVERT": lambda self: self._parse_convert(False), "EXTRACT": lambda self: self._parse_extract(), "POSITION": lambda self: self._parse_position(), + "STRING_AGG": lambda self: self._parse_string_agg(), "SUBSTRING": lambda self: self._parse_substring(), "TRIM": lambda self: self._parse_trim(), - "CAST": lambda self: self._parse_cast(self.STRICT_CAST), "TRY_CAST": lambda self: self._parse_cast(False), - "STRING_AGG": lambda self: self._parse_string_agg(), + "TRY_CONVERT": lambda self: self._parse_convert(False), } QUERY_MODIFIER_PARSERS = { @@ -684,13 +687,28 @@ class Parser(metaclass=_Parser): "sample": lambda self: self._parse_table_sample(as_modifier=True), } + SET_PARSERS = { + "GLOBAL": lambda self: self._parse_set_item_assignment("GLOBAL"), + "LOCAL": lambda self: self._parse_set_item_assignment("LOCAL"), + "SESSION": lambda self: self._parse_set_item_assignment("SESSION"), + "TRANSACTION": lambda self: self._parse_set_transaction(), + } + SHOW_PARSERS: t.Dict[str, t.Callable] = {} - SET_PARSERS: t.Dict[str, t.Callable] = {} MODIFIABLES = (exp.Subquery, exp.Subqueryable, exp.Table) TRANSACTION_KIND = {"DEFERRED", "IMMEDIATE", "EXCLUSIVE"} + TRANSACTION_CHARACTERISTICS = { + "ISOLATION LEVEL REPEATABLE READ", + "ISOLATION LEVEL READ COMMITTED", + "ISOLATION LEVEL READ UNCOMMITTED", + "ISOLATION LEVEL SERIALIZABLE", + "READ WRITE", + "READ ONLY", + } + INSERT_ALTERNATIVES = {"ABORT", "FAIL", "IGNORE", "REPLACE", "ROLLBACK"} WINDOW_ALIAS_TOKENS = ID_VAR_TOKENS - {TokenType.ROWS} @@ -1775,11 +1793,12 @@ class Parser(metaclass=_Parser): self, alias_tokens: t.Optional[t.Collection[TokenType]] = None ) -> t.Optional[exp.Expression]: any_token = self._match(TokenType.ALIAS) - alias = self._parse_id_var( - any_token=any_token, tokens=alias_tokens or self.TABLE_ALIAS_TOKENS + alias = ( + self._parse_id_var(any_token=any_token, tokens=alias_tokens or self.TABLE_ALIAS_TOKENS) + or self._parse_string_as_identifier() ) - index = self._index + index = self._index if self._match(TokenType.L_PAREN): columns = self._parse_csv(self._parse_function_parameter) self._match_r_paren() if columns else self._retreat(index) @@ -2046,7 +2065,12 @@ class Parser(metaclass=_Parser): def _parse_table_parts(self, schema: bool = False) -> exp.Expression: catalog = None db = None - table = (not schema and self._parse_function()) or self._parse_id_var(any_token=False) + + table = ( + (not schema and self._parse_function()) + or self._parse_id_var(any_token=False) + or self._parse_string_as_identifier() + ) while self._match(TokenType.DOT): if catalog: @@ -2085,6 +2109,8 @@ class Parser(metaclass=_Parser): subquery = self._parse_select(table=True) if subquery: + if not subquery.args.get("pivots"): + subquery.set("pivots", self._parse_pivots()) return subquery this = self._parse_table_parts(schema=schema) @@ -3370,9 +3396,9 @@ class Parser(metaclass=_Parser): def _parse_window( self, this: t.Optional[exp.Expression], alias: bool = False ) -> t.Optional[exp.Expression]: - if self._match(TokenType.FILTER): - where = self._parse_wrapped(self._parse_where) - this = self.expression(exp.Filter, this=this, expression=where) + if self._match_pair(TokenType.FILTER, TokenType.L_PAREN): + this = self.expression(exp.Filter, this=this, expression=self._parse_where()) + self._match_r_paren() # T-SQL allows the OVER (...) syntax after WITHIN GROUP. # https://learn.microsoft.com/en-us/sql/t-sql/functions/percentile-disc-transact-sql?view=sql-server-ver16 @@ -3504,6 +3530,9 @@ class Parser(metaclass=_Parser): return self.PRIMARY_PARSERS[TokenType.STRING](self, self._prev) return self._parse_placeholder() + def _parse_string_as_identifier(self) -> t.Optional[exp.Expression]: + return exp.to_identifier(self._match(TokenType.STRING) and self._prev.text, quoted=True) + def _parse_number(self) -> t.Optional[exp.Expression]: if self._match(TokenType.NUMBER): return self.PRIMARY_PARSERS[TokenType.NUMBER](self, self._prev) @@ -3778,23 +3807,6 @@ class Parser(metaclass=_Parser): ) return self._parse_as_command(start) - def _parse_show(self) -> t.Optional[exp.Expression]: - parser = self._find_parser(self.SHOW_PARSERS, self._show_trie) # type: ignore - if parser: - return parser(self) - self._advance() - return self.expression(exp.Show, this=self._prev.text.upper()) - - def _default_parse_set_item(self) -> exp.Expression: - return self.expression( - exp.SetItem, - this=self._parse_statement(), - ) - - def _parse_set_item(self) -> t.Optional[exp.Expression]: - parser = self._find_parser(self.SET_PARSERS, self._set_trie) # type: ignore - return parser(self) if parser else self._default_parse_set_item() - def _parse_merge(self) -> exp.Expression: self._match(TokenType.INTO) target = self._parse_table() @@ -3861,8 +3873,71 @@ class Parser(metaclass=_Parser): expressions=whens, ) + def _parse_show(self) -> t.Optional[exp.Expression]: + parser = self._find_parser(self.SHOW_PARSERS, self._show_trie) # type: ignore + if parser: + return parser(self) + self._advance() + return self.expression(exp.Show, this=self._prev.text.upper()) + + def _parse_set_item_assignment( + self, kind: t.Optional[str] = None + ) -> t.Optional[exp.Expression]: + index = self._index + + if kind in {"GLOBAL", "SESSION"} and self._match_text_seq("TRANSACTION"): + return self._parse_set_transaction(global_=kind == "GLOBAL") + + left = self._parse_primary() or self._parse_id_var() + + if not self._match_texts(("=", "TO")): + self._retreat(index) + return None + + right = self._parse_statement() or self._parse_id_var() + this = self.expression( + exp.EQ, + this=left, + expression=right, + ) + + return self.expression( + exp.SetItem, + this=this, + kind=kind, + ) + + def _parse_set_transaction(self, global_: bool = False) -> exp.Expression: + self._match_text_seq("TRANSACTION") + characteristics = self._parse_csv( + lambda: self._parse_var_from_options(self.TRANSACTION_CHARACTERISTICS) + ) + return self.expression( + exp.SetItem, + expressions=characteristics, + kind="TRANSACTION", + **{"global": global_}, # type: ignore + ) + + def _parse_set_item(self) -> t.Optional[exp.Expression]: + parser = self._find_parser(self.SET_PARSERS, self._set_trie) # type: ignore + return parser(self) if parser else self._parse_set_item_assignment(kind=None) + def _parse_set(self) -> exp.Expression: - return self.expression(exp.Set, expressions=self._parse_csv(self._parse_set_item)) + index = self._index + set_ = self.expression(exp.Set, expressions=self._parse_csv(self._parse_set_item)) + + if self._curr: + self._retreat(index) + return self._parse_as_command(self._prev) + + return set_ + + def _parse_var_from_options(self, options: t.Collection[str]) -> t.Optional[exp.Expression]: + for option in options: + if self._match_text_seq(*option.split(" ")): + return exp.Var(this=option) + return None def _parse_as_command(self, start: Token) -> exp.Command: while self._curr: @@ -3874,6 +3949,9 @@ class Parser(metaclass=_Parser): def _find_parser( self, parsers: t.Dict[str, t.Callable], trie: t.Dict ) -> t.Optional[t.Callable]: + if not self._curr: + return None + index = self._index this = [] while True: @@ -3973,7 +4051,16 @@ class Parser(metaclass=_Parser): return this def _replace_lambda(self, node, lambda_variables): - if isinstance(node, exp.Column): - if node.name in lambda_variables: - return node.this + for column in node.find_all(exp.Column): + if column.parts[0].name in lambda_variables: + dot_or_id = column.to_dot() if column.table else column.this + parent = column.parent + + while isinstance(parent, exp.Dot): + if not isinstance(parent.parent, exp.Dot): + parent.replace(dot_or_id) + break + parent = parent.parent + else: + column.replace(dot_or_id) return node |