fix(rust, python): block predicate pushdown is_in and null producing … (

pola-rs#10194)
cmdlineluser · Jul 31, 2023 · 47b91ab · 47b91ab
1 parent 600c325
commit 47b91ab
Show file tree

Hide file tree

Showing 2 changed files with 20 additions and 0 deletions.
diff --git a/crates/polars-plan/src/logical_plan/optimizer/predicate_pushdown/join.rs b/crates/polars-plan/src/logical_plan/optimizer/predicate_pushdown/join.rs
@@ -11,6 +11,11 @@ fn should_block_join_specific(ae: &AExpr, how: &JoinType) -> bool {
                 | FunctionExpr::FillNull { .. },
             ..
         } => join_produces_null(how),
+        #[cfg(feature = "is_in")]
+        Function {
+            function: FunctionExpr::Boolean(BooleanFunction::IsIn),
+            ..
+        } => join_produces_null(how),
         // joins can produce duplicates
         #[cfg(feature = "is_unique")]
         Function {

diff --git a/py-polars/tests/unit/test_predicates.py b/py-polars/tests/unit/test_predicates.py
@@ -158,3 +158,18 @@ def test_predicate_pushdown_join_fill_null_10058() -> None:
         .collect()
         .to_dict(False)["id"]
     ) == [0, 2]
+
+
+def test_is_in_join_blocked() -> None:
+    df1 = pl.DataFrame(
+        {"Groups": ["A", "B", "C", "D", "E", "F"], "values0": [1, 2, 3, 4, 5, 6]}
+    ).lazy()
+
+    df2 = pl.DataFrame(
+        {"values22": [1, 2, None, 4, 5, 6], "values20": [1, 2, 3, 4, 5, 6]}
+    ).lazy()
+
+    df_all = df2.join(df1, left_on="values20", right_on="values0", how="left")
+    assert df_all.filter(~pl.col("Groups").is_in(["A", "B", "F"])).collect().to_dict(
+        False
+    ) == {"values22": [None, 4, 5], "values20": [3, 4, 5], "Groups": ["C", "D", "E"]}