fix: only do row identity based joins when joining by index

bigframes/core/__init__.py

@@ -349,7 +349,7 @@ def join(
         self,
         other: ArrayValue,
         join_def: join_def.JoinDefinition,
-        allow_row_identity_join: bool = True,
+        allow_row_identity_join: bool = False,
     ):
         return ArrayValue(
             nodes.JoinNode(

bigframes/core/compile/single_column.py

@@ -33,7 +33,7 @@ def join_by_column_ordered(
     left: compiled.OrderedIR,
     right: compiled.OrderedIR,
     join: join_defs.JoinDefinition,
-    allow_row_identity_join: bool = True,
+    allow_row_identity_join: bool = False,
 ) -> compiled.OrderedIR:
     """Join two expressions by column equality.
 
@@ -134,7 +134,7 @@ def join_by_column_unordered(
     left: compiled.UnorderedIR,
     right: compiled.UnorderedIR,
     join: join_defs.JoinDefinition,
-    allow_row_identity_join: bool = True,
+    allow_row_identity_join: bool = False,
 ) -> compiled.UnorderedIR:
     """Join two expressions by column equality.
 

bigframes/core/nodes.py

@@ -115,7 +115,7 @@ class JoinNode(BigFrameNode):
     left_child: BigFrameNode
     right_child: BigFrameNode
     join: JoinDefinition
-    allow_row_identity_join: bool = True
+    allow_row_identity_join: bool = False
 
     @property
     def row_preserving(self) -> bool:

tests/system/conftest.py

@@ -285,6 +285,13 @@ def scalars_table_id(test_data_tables) -> str:
     return test_data_tables["scalars"]
 
 
+@pytest.fixture(scope="session")
+def baseball_schedules_df(session: bigframes.Session) -> bigframes.dataframe.DataFrame:
+    """Public BQ table"""
+    df = session.read_gbq("bigquery-public-data.baseball.schedules")
+    return df
+
+
 @pytest.fixture(scope="session")
 def hockey_table_id(test_data_tables) -> str:
     return test_data_tables["hockey_players"]

tests/system/small/regression/test_issue355_merge_after_filter.py

@@ -0,0 +1,70 @@
+# Copyright 2024 Google LLC
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import pandas as pd
+import pytest
+
+from tests.system.utils import assert_pandas_df_equal
+
+
+@pytest.mark.parametrize(
+    ("merge_how",),
+    [
+        ("inner",),
+        ("outer",),
+        ("left",),
+        ("right",),
+    ],
+)
+def test_merge_after_filter(baseball_schedules_df, merge_how):
+    on = ["awayTeamName"]
+    left_columns = [
+        "gameId",
+        "year",
+        "homeTeamName",
+        "awayTeamName",
+        "duration_minutes",
+    ]
+    right_columns = [
+        "gameId",
+        "year",
+        "homeTeamName",
+        "awayTeamName",
+        "duration_minutes",
+    ]
+
+    left = baseball_schedules_df[left_columns]
+    left = left[left["homeTeamName"] == "Rays"]
+    # Offset the rows somewhat so that outer join can have an effect.
+    right = baseball_schedules_df[right_columns]
+    right = right[right["homeTeamName"] == "White Sox"]
+
+    df = left.merge(right, on=on, how=merge_how)
+    bf_result = df.to_pandas()
+
+    left_pandas = baseball_schedules_df.to_pandas()[left_columns]
+    left_pandas = left_pandas[left_pandas["homeTeamName"] == "Rays"]
+
+    right_pandas = baseball_schedules_df.to_pandas()[right_columns]
+    right_pandas = right_pandas[right_pandas["homeTeamName"] == "White Sox"]
+
+    pd_result = pd.merge(
+        left_pandas,
+        right_pandas,
+        merge_how,
+        on,
+        sort=True,
+    )
+
+    assert_pandas_df_equal(bf_result, pd_result, ignore_order=True)