Github user paul-rogers commented on a diff in the pull request:

    https://github.com/apache/drill/pull/938#discussion_r137939168
  
    --- Diff: 
exec/java-exec/src/main/java/org/apache/drill/exec/physical/impl/aggregate/HashAggBatch.java
 ---
    @@ -293,7 +299,7 @@ private HashAggregator createAggregatorInternal() 
throws SchemaChangeException,
             aggrExprs,
             cgInner.getWorkspaceTypes(),
             groupByOutFieldIds,
    -        this.container);
    +        this.container, extraNonNullColumns * 8 /* sizeof(BigInt) */);
    --- End diff --
    
    If the `BigInt` column is used to indicate nulls, then each value is of 
size 9. And, since, on average, each vector has 25% internal fragmentation. To 
account for this, perhaps assume that the average size is 12 or 13 bytes.


---

Reply via email to