[
https://issues.apache.org/jira/browse/PHOENIX-2724?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15262067#comment-15262067
]
Ankit Singhal commented on PHOENIX-2724:
----------------------------------------
[~samarthjain], are we not ignoring serial hint if query in serial is not
possible?
In ScanPlan:-
{code}
private static boolean isSerial(StatementContext context, FilterableStatement
statement,
TableRef tableRef, OrderBy orderBy, Integer limit, Integer offset,
boolean allowPageFilter) throws SQLException {
if (statement.getHint().hasHint(HintNode.Hint.SERIAL)) {
return true;
}
{code}
And in AggregatePlan:-
{code}
BaseResultIterators iterators =
statement.getHint().hasHint(HintNode.Hint.SERIAL)
? new SerialIterators(this, null, null,
wrapParallelIteratorFactory(), scanGrouper)
: new ParallelIterators(this, null,
wrapParallelIteratorFactory());
{code}
> Query with large number of guideposts is slower compared to no stats
> --------------------------------------------------------------------
>
> Key: PHOENIX-2724
> URL: https://issues.apache.org/jira/browse/PHOENIX-2724
> Project: Phoenix
> Issue Type: Bug
> Affects Versions: 4.7.0
> Environment: Phoenix 4.7.0-RC4, HBase-0.98.17 on a 8 node cluster
> Reporter: Mujtaba Chohan
> Assignee: Samarth Jain
> Fix For: 4.8.0
>
> Attachments: PHOENIX-2724.patch, PHOENIX-2724_v2.patch
>
>
> With 1MB guidepost width for ~900GB/500M rows table. Queries with short scan
> range gets significantly slower.
> Without stats:
> {code}
> select * from T limit 10; // query execution time <100 msec
> {code}
> With stats:
> {code}
> select * from T limit 10; // query execution time >20 seconds
> Explain plan: CLIENT 876085-CHUNK 476569382 ROWS 876060986727 BYTES SERIAL
> 1-WAY FULL SCAN OVER T SERVER 10 ROW LIMIT CLIENT 10 ROW LIMIT
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
