Michael Smith has uploaded this change for review. ( 
http://gerrit.cloudera.org:8080/19898


Change subject: IMPALA-10186: Fix writing empty parquet page
......................................................................

IMPALA-10186: Fix writing empty parquet page

Fixeus writing an empty parquet page when a page fills at the same time
that its dictionary fills. We had logic to detect both, but they were
handled separately and could trigger sequentially, resulting in creating
two new pages in a row. Now if a page becomes full, we also check
whether the dictionary is full and flush it if needed.

Dictionary size is checked after insert fails because if the item
already exists in the dictionary then we don't need a new page.

Testing: manually tested by copying a table with an empty page.

Change-Id: I90d30d958f07c6289a1beba1b5df1ab3d7213799
---
M be/src/exec/parquet/hdfs-parquet-table-writer.cc
M be/src/util/dict-encoding.h
2 files changed, 13 insertions(+), 3 deletions(-)



  git pull ssh://gerrit.cloudera.org:29418/Impala-ASF refs/changes/98/19898/1
--
To view, visit http://gerrit.cloudera.org:8080/19898
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: newchange
Gerrit-Change-Id: I90d30d958f07c6289a1beba1b5df1ab3d7213799
Gerrit-Change-Number: 19898
Gerrit-PatchSet: 1
Gerrit-Owner: Michael Smith <[email protected]>

Reply via email to