Fix: left anti join filter empty rows.

apache · Apr 15, 2024 · 6bdcea8 · 6bdcea8
1 parent a6cdcd0
commit 6bdcea8
Show file tree

Hide file tree

Showing 2 changed files with 22 additions and 0 deletions.
diff --git a/cpp/src/arrow/acero/hash_join_node_test.cc b/cpp/src/arrow/acero/hash_join_node_test.cc
@@ -2036,6 +2036,25 @@ TEST(HashJoin, ResidualFilter) {
                   [3, 4, "alpha", 4, 16, "alpha"]])")});
 }
 
+TEST(HashJoin, FilterEmptyRows) {
+  BatchesWithSchema input_left;
+  input_left.batches = {ExecBatchFromJSON({int32(), utf8(), int32()}, R"([[2, "Jarry", 28]])")};
+  input_left.schema = schema({field("id", int32()), field("name", utf8()), field("age", int32())});
+
+  BatchesWithSchema input_right;
+  input_right.batches = {ExecBatchFromJSON(
+      {int32(), int32(), utf8()},
+      R"([[2, 10, "Jack"], [3, 12, "Mark"], [4, 15, "Tom"], [1, 10, "Jack"]])")};
+  input_right.schema = schema({field("id", int32()), field("stu_id", int32()), field("subject", utf8())});
+
+  const ResidualFilterCaseRunner runner{std::move(input_left), std::move(input_right)};
+
+  Expression filter = greater(field_ref("age"), literal(25));
+
+  runner.Run(JoinType::LEFT_ANTI, {"id"}, {"stu_id"}, std::move(filter),
+             {ExecBatchFromJSON({int32(), utf8(), int32()}, R"([[2, "Jarry", 28]])")});
+}
+
 TEST(HashJoin, TrivialResidualFilter) {
   Expression always_true =
       equal(call("add", {field_ref("l1"), field_ref("r1")}), literal(2));  // 1 + 1 == 2

diff --git a/cpp/src/arrow/acero/swiss_join.cc b/cpp/src/arrow/acero/swiss_join.cc
@@ -2160,6 +2160,9 @@ Status JoinResidualFilter::FilterOneBatch(const ExecBatch& keypayload_batch,
                                           bool output_key_ids, bool output_payload_ids,
                                           arrow::util::TempVectorStack* temp_stack,
                                           int* num_passing_rows) const {
+  if (num_batch_rows == 0) {
+    return Status::OK();
+  }
   // Caller must do shortcuts for trivial filter.
   ARROW_DCHECK(!filter_.IsNullLiteral() && filter_ != literal(true) &&
                filter_ != literal(false));