Github user maropu commented on the pull request:
https://github.com/apache/spark/pull/11236#issuecomment-185117901
@nongli This discussion comes from #10965;
Before I make pr to improve compression performance of columnar caches, I
have some questions about `InMemoryRelation`. In the current codes of
`InMemoryRelation`, `ColumnBuilder` buffers input tuples in heap space and
compresses them in bulk. Do you have any plan to use off-heap for this
compression processing in `ColumnBuilder`? IMO we can fix this by adding some
functionality in `ColumnVector` you implemented for parquet vectorized
decoding, and then by using the extended `ColumnVector` to buffer input tuples
inside `ColumnBuilder`. What do you think?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
