Github user manishnalla1994 commented on the issue:
https://github.com/apache/carbondata/pull/3056
@xuchuanyin Datasource table uses direct filling flow. As in direct flow
there is no intermediate buffer so we are not using off-heap to store the page
data in memory(filling all the records of a page to vector instead of filling
batch wise). So in this case we can remove freeing of unsafe memory for Query
as its not required.
In case of stored by table, handling will be different as we support both
batch wise filling and direct filling and for batch filling we are using
unsafe, so we have to clear unsafe memory in this case.
Here same handling is not required for data source table.
Please refer https://github.com/apache/carbondata/pull/2591 for stored by
handling of this issue.
---
