Adding upstream version 25.1.0.upstream/25.1.0

Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>

1 files changed, 27 insertions, 10 deletions

diff --git a/sqlglot/parser.py b/sqlglot/parser.py
index ed53b19..cf5cc9d 100644
--- a/sqlglot/parser.py
+++ b/sqlglot/parser.py
@@ -588,11 +588,12 @@ class Parser(metaclass=_Parser):
     }
 
     JOIN_KINDS = {
+        TokenType.ANTI,
+        TokenType.CROSS,
         TokenType.INNER,
         TokenType.OUTER,
-        TokenType.CROSS,
         TokenType.SEMI,
-        TokenType.ANTI,
+        TokenType.STRAIGHT_JOIN,
     }
 
     JOIN_HINTS: t.Set[str] = set()
@@ -1065,7 +1066,7 @@ class Parser(metaclass=_Parser):
         exp.DataType.Type.JSON: lambda self, this, _: self.expression(exp.ParseJSON, this=this),
     }
 
-    TYPE_CONVERTER: t.Dict[exp.DataType.Type, t.Callable[[exp.DataType], exp.DataType]] = {}
+    TYPE_CONVERTERS: t.Dict[exp.DataType.Type, t.Callable[[exp.DataType], exp.DataType]] = {}
 
     DDL_SELECT_TOKENS = {TokenType.SELECT, TokenType.WITH, TokenType.L_PAREN}
 
@@ -1138,7 +1139,14 @@ class Parser(metaclass=_Parser):
 
     FETCH_TOKENS = ID_VAR_TOKENS - {TokenType.ROW, TokenType.ROWS, TokenType.PERCENT}
 
-    ADD_CONSTRAINT_TOKENS = {TokenType.CONSTRAINT, TokenType.PRIMARY_KEY, TokenType.FOREIGN_KEY}
+    ADD_CONSTRAINT_TOKENS = {
+        TokenType.CONSTRAINT,
+        TokenType.FOREIGN_KEY,
+        TokenType.INDEX,
+        TokenType.KEY,
+        TokenType.PRIMARY_KEY,
+        TokenType.UNIQUE,
+    }
 
     DISTINCT_TOKENS = {TokenType.DISTINCT}
 
@@ -3099,7 +3107,7 @@ class Parser(metaclass=_Parser):
         index = self._index
         method, side, kind = self._parse_join_parts()
         hint = self._prev.text if self._match_texts(self.JOIN_HINTS) else None
-        join = self._match(TokenType.JOIN)
+        join = self._match(TokenType.JOIN) or (kind and kind.token_type == TokenType.STRAIGHT_JOIN)
 
         if not skip_join_token and not join:
             self._retreat(index)
@@ -3242,7 +3250,7 @@ class Parser(metaclass=_Parser):
             while self._match_set(self.TABLE_INDEX_HINT_TOKENS):
                 hint = exp.IndexTableHint(this=self._prev.text.upper())
 
-                self._match_texts(("INDEX", "KEY"))
+                self._match_set((TokenType.INDEX, TokenType.KEY))
                 if self._match(TokenType.FOR):
                     hint.set("target", self._advance_any() and self._prev.text.upper())
 
@@ -4464,8 +4472,8 @@ class Parser(metaclass=_Parser):
             )
             self._match(TokenType.R_BRACKET)
 
-        if self.TYPE_CONVERTER and isinstance(this.this, exp.DataType.Type):
-            converter = self.TYPE_CONVERTER.get(this.this)
+        if self.TYPE_CONVERTERS and isinstance(this.this, exp.DataType.Type):
+            converter = self.TYPE_CONVERTERS.get(this.this)
             if converter:
                 this = converter(t.cast(exp.DataType, this))
 
@@ -4496,7 +4504,12 @@ class Parser(metaclass=_Parser):
 
     def _parse_column(self) -> t.Optional[exp.Expression]:
         this = self._parse_column_reference()
-        return self._parse_column_ops(this) if this else self._parse_bracket(this)
+        column = self._parse_column_ops(this) if this else self._parse_bracket(this)
+
+        if self.dialect.SUPPORTS_COLUMN_JOIN_MARKS and column:
+            column.set("join_mark", self._match(TokenType.JOIN_MARKER))
+
+        return column
 
     def _parse_column_reference(self) -> t.Optional[exp.Expression]:
         this = self._parse_field()
@@ -4522,7 +4535,11 @@ class Parser(metaclass=_Parser):
 
         while self._match(TokenType.COLON):
             start_index = self._index
-            path = self._parse_column_ops(self._parse_field(any_token=True))
+
+            # Snowflake allows reserved keywords as json keys but advance_any() excludes TokenType.SELECT from any_tokens=True
+            path = self._parse_column_ops(
+                self._parse_field(any_token=True, tokens=(TokenType.SELECT,))
+            )
 
             # The cast :: operator has a lower precedence than the extraction operator :, so
             # we rearrange the AST appropriately to avoid casting the JSON path