[
https://issues.apache.org/jira/browse/HIVE-7990?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Mostafa Mokhtar updated HIVE-7990:
----------------------------------
Description: For queries with rollup and cube (was: When loading into an
un-paritioned ORC table WriterImpl$StructTreeWriter.write method is
synchronized.
When hive.optimize.sort.dynamic.partition is enabled the current thread will be
the only writer and the synchronization is not needed.
Also checking for memory per row is an over kill , this can be done per 1K
rows or such
{code}
public void addRow(Object row) throws IOException {
synchronized (this) {
treeWriter.write(row);
rowsInStripe += 1;
if (buildIndex) {
rowsInIndex += 1;
if (rowsInIndex >= rowIndexStride) {
createRowIndexEntry();
}
}
}
memoryManager.addedRow();
}
{code}
This can improve ORC load performance by 7%
{code}
Stack Trace Sample Count Percentage(%)
WriterImpl.addRow(Object) 5,852 65.782
WriterImpl$StructTreeWriter.write(Object) 5,163 58.037
MemoryManager.addedRow() 666 7.487
MemoryManager.notifyWriters() 648 7.284
WriterImpl.checkMemory(double) 645 7.25
WriterImpl.flushStripe() 643 7.228
WriterImpl$StructTreeWriter.writeStripe(OrcProto$StripeFooter$Builder, int)
584 6.565
{code}
)
> With fetch column stats disabled number of elements in grouping set is not
> taken into account
> ---------------------------------------------------------------------------------------------
>
> Key: HIVE-7990
> URL: https://issues.apache.org/jira/browse/HIVE-7990
> Project: Hive
> Issue Type: Bug
> Components: Statistics
> Affects Versions: 0.13.1
> Reporter: Mostafa Mokhtar
> Assignee: Prasanth J
> Labels: performance
> Fix For: 0.14.0
>
>
> For queries with rollup and cube
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
