Github user paul-rogers commented on a diff in the pull request:
https://github.com/apache/drill/pull/1059#discussion_r155939506
--- Diff:
exec/java-exec/src/test/java/org/apache/drill/exec/physical/unit/TestNullInputMiniPlan.java
---
@@ -323,8 +323,6 @@ public void testHashJoinLeftEmpty() throws Exception {
.build();
BatchSchema expectedSchema = new SchemaBuilder()
- .addNullable("a", TypeProtos.MinorType.BIGINT)
- .addNullable("b", TypeProtos.MinorType.BIGINT)
--- End diff --
I wonder about this. This test came out of the "empty batches" project
which attempted to handle empty inputs. If this PR finds the need to change the
tests, then we are changing semantics of how we handle schemas. Are we sure we
want to make this change?
If we are changing the schema, are we saying that if we get no left input
(but do get a schema), that we won't product the joined schema? This would seem
to be a bug rather than a feature.
---
