You are viewing a plain text version of this content. The canonical link for it is here.

Posted to user@hive.apache.org by Bowen Zhang <bo...@stripe.com> on 2020/03/09 22:26:44 UTC

Presto query planning is slow with Hive 4.0

Hi,

I have a Presto (322) + Hive (2.3.4) + S3 setup that works fine. A simple
"select count(*) from mydb.mytable" (underneath data is a single 52KB
parquet file on S3) takes 200ms to complete.

I needed fix for https://issues.apache.org/jira/browse/HIVE-21631 so I
dropped in a locally built (from HEAD of master) Hive 4.0. No change to
config under /etc/hive. Now the same query took 20s to complete.
Attached is log from hive-metastore.log during the execution of the query.

Any suggestions on direction of investigation?

-Bowen