Tim Armstrong has posted comments on this change. Change subject: IMPALA-4846: Upgrade Snappy to 1.1.4 ......................................................................
Patch Set 1: Can you do a quick benchmark to sanity-check the performance scanning compressed parquet? You can do something like this: use tpch_parquet; create table biglineitem stored as parquet as select * from lineitem; insert into biglineitem select * from biglineitem; -- Repeat a few times to get a table that takes a few seconds to scan -- Use one thread and one impala daemon to reduce variability set mt_dop=1; set num_nodes=1; select * from biglineitem where l_orderkey = 3; When running that locally I see snappy taking ~18% of CPU in perf top so hopefully the new snappy version will shave some time off the query. -- To view, visit http://gerrit.cloudera.org:8080/6428 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I2f3439019ae22bbcf4db7f731e45ba5f7899fcc2 Gerrit-PatchSet: 1 Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-Owner: Laszlo Gaal <[email protected]> Gerrit-Reviewer: Tim Armstrong <[email protected]> Gerrit-HasComments: No
