Fix: performance regression due to 6f80cc8

tobymao · Jul 4, 2023 · d1ad7da · d1ad7da
1 parent 5a84605
commit d1ad7da
Show file tree

Hide file tree

Showing 2 changed files with 61 additions and 12 deletions.
diff --git a/sqlglot/parser.py b/sqlglot/parser.py
@@ -2277,23 +2277,22 @@ def _parse_join(self, skip_join_token: bool = False) -> t.Optional[exp.Join]:
         if hint:
             kwargs["hint"] = hint
 
-        def parse_join_condition(kwargs: t.Dict) -> None:
-            if self._match(TokenType.ON):
-                kwargs["on"] = self._parse_conjunction()
-            elif self._match(TokenType.USING):
-                kwargs["using"] = self._parse_wrapped_id_vars()
-
-        parse_join_condition(kwargs)
-
-        if not (kind and kind.token_type == TokenType.CROSS):
+        if self._match(TokenType.ON):
+            kwargs["on"] = self._parse_conjunction()
+        elif self._match(TokenType.USING):
+            kwargs["using"] = self._parse_wrapped_id_vars()
+        elif not (kind and kind.token_type == TokenType.CROSS):
             index = self._index
             joins = self._parse_joins()
 
-            if joins and self._match_set({TokenType.ON, TokenType.USING}, advance=False):
-                parse_join_condition(kwargs)
-                kwargs["this"].set("joins", joins)
+            if joins and self._match(TokenType.ON):
+                kwargs["on"] = self._parse_conjunction()
+            elif joins and self._match(TokenType.USING):
+                kwargs["using"] = self._parse_wrapped_id_vars()
             else:
+                joins = None
                 self._retreat(index)
+            kwargs["this"].set("joins", joins)
 
         return self.expression(exp.Join, **kwargs)
 

diff --git a/tests/test_parser.py b/tests/test_parser.py
@@ -1,3 +1,4 @@
+import time
 import unittest
 from unittest.mock import patch
 
@@ -523,6 +524,55 @@ def test_pivot_columns(self):
                 columns = expr.args["from"].this.args["pivots"][0].args["columns"]
                 self.assertEqual(expected_columns, [col.sql(dialect=dialect) for col in columns])
 
+    def test_parse_nested(self):
+        now = time.time()
+        query = parse_one(
+            """
+            select *
+            FROM a
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            LEFT JOIN b ON a.id = b.id
+            """
+        )
+        self.assertIsNotNone(query)
+        self.assertLessEqual(time.time() - now, 0.1)
+
     def test_parse_properties(self):
         self.assertEqual(
             parse_one("create materialized table x").sql(), "CREATE MATERIALIZED TABLE x"