Github user paul-rogers commented on a diff in the pull request: https://github.com/apache/drill/pull/1101#discussion_r164623893 --- Diff: exec/java-exec/src/main/java/org/apache/drill/exec/physical/impl/spill/RecordBatchSizer.java --- @@ -129,11 +143,16 @@ public ColumnSize(ValueVector v, String prefix) { // No standard size for Union type dataSize = v.getPayloadByteCount(valueCount); break; + case GENERIC_OBJECT: + // We cannot provide a size for Generic Objects --- End diff -- The `GENERIC_OBJECT` type is used any time we do a system table query: system tables are represented as Java objects. There is an open question about certain aggregate functions. (See a note sent a week or so ago.) These aggregate functions use an `ObjectHolder` as their `\@Workspace`. @Ben-Zvi and I discussed whether such aggregates are spillable. This may be an unresolved issue.
---