[
https://issues.apache.org/jira/browse/HIVE-20621?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16624409#comment-16624409
]
Gopal V commented on HIVE-20621:
--------------------------------
LGTM - +1
{code} {
operationStatus = ((HiveStatement) statement).waitForOperationToComplete();
{code}
add a comment saying that a query is expected to go from running -> complete
and won't go back to running while fetching results (this is an implicit state
machine).
> GetOperationStatus called in resultset.next causing incremental slowness
> ------------------------------------------------------------------------
>
> Key: HIVE-20621
> URL: https://issues.apache.org/jira/browse/HIVE-20621
> Project: Hive
> Issue Type: Bug
> Components: JDBC
> Affects Versions: 4.0.0, 3.2.0
> Reporter: Prasanth Jayachandran
> Assignee: Prasanth Jayachandran
> Priority: Major
> Attachments: HIVE-20621.1.patch
>
>
> Fetching result set for a result cache hit query gets slower as more rows are
> fetched. For fetching 10 row result set it took about 900ms but fetching 200
> row result set took 8 seconds.
> Reason for this slowness is GetOperationStatus is invoked inside
> resultset.next() and it happens for every row even after operation has
> completed. This is one RPC call per row fetched (there is also connection
> overhead without keepalive).
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
