Github user ppadma commented on a diff in the pull request:
https://github.com/apache/drill/pull/1107#discussion_r167380394
--- Diff:
exec/java-exec/src/main/java/org/apache/drill/exec/physical/impl/spill/RecordBatchSizer.java
---
@@ -137,10 +137,10 @@ public ColumnSize(ValueVector v, String prefix) {
case MAP:
case UNION:
// No standard size for Union type
- dataSize = v.getPayloadByteCount(valueCount);
+ dataSize = valueCount == 0 ? 0 : v.getPayloadByteCount(valueCount);
--- End diff --
When we get empty batch, getPayloadByteCount will throw out of bounds
exception if the underlying method tries to read the value vector for index 0
i.e. when it is trying to read offset vector. I added the check in those
methods and removed here.
I still have to retain the check here for the case when we are trying to
read the offset vector.
---
