Alex Behm has posted comments on this change.

Change subject: Optimized ReadValueBatch() for Parquet scalar column readers.
......................................................................


Patch Set 5:

(1 comment)

http://gerrit.cloudera.org:8080/#/c/2843/5/be/src/exec/hdfs-parquet-scanner.cc
File be/src/exec/hdfs-parquet-scanner.cc:

Line 827:   bool MaterializeValueBatch(MemPool* pool, int max_values, int 
tuple_size,
> That logic is already specialized and in such cases we will stamp out row b
Unfortunately, what I said was only almost accurate. There is one special case 
that will still hit this code path with !MATERIALIZED && !IN_COLLECTION. It can 
happen if we have a query that only references columns that fail to resolve in 
the data files.

As Marcel suggested, I hoisted out that part into this function directly. The 
change is simple but please have another look.

I'm doing another exhaustive run.


-- 
To view, visit http://gerrit.cloudera.org:8080/2843
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: I21fa9b050a45f2dd45cc0091ea5b008d3c0a3f30
Gerrit-PatchSet: 5
Gerrit-Project: Impala
Gerrit-Branch: cdh5-trunk
Gerrit-Owner: Alex Behm <[email protected]>
Gerrit-Reviewer: Alex Behm <[email protected]>
Gerrit-Reviewer: Internal Jenkins
Gerrit-Reviewer: Marcel Kornacker <[email protected]>
Gerrit-Reviewer: Mostafa Mokhtar <[email protected]>
Gerrit-Reviewer: Tim Armstrong <[email protected]>
Gerrit-HasComments: Yes

Reply via email to