[
https://issues.apache.org/jira/browse/PHOENIX-3005?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15338870#comment-15338870
]
Lars Hofhansl commented on PHOENIX-3005:
----------------------------------------
Is this change safe?
{code}
- if (isOrderPreserving) {
- return new
GroupBy.GroupByBuilder(this).setOrderPreservingColumnCount(orderPreservingColumnCount).build();
- }
- if (isUngroupedAggregate) {
- return UNGROUPED_GROUP_BY;
+ if (isOrderPreserving || isUngroupedAggregate) {
+ return new
GroupBy.GroupByBuilder(this).setIsOrderPreserving(isOrderPreserving).setOrderPreservingColumnCount(orderPreservingColumnCount).build();
}
{code}
So if this is ungrouped aggregate and we'd return the a new group by (rather
than UNGROUPED_GROUP_BY) even when order preserving is false. (This change is
needed BTW, because UNGROUPED_GROUP_BY returns true from isOrderPreserving)
Looks like it's safe.
> Use DistinctPrefixFilter for COUNT(DISTINCT...) index scans
> -----------------------------------------------------------
>
> Key: PHOENIX-3005
> URL: https://issues.apache.org/jira/browse/PHOENIX-3005
> Project: Phoenix
> Issue Type: Sub-task
> Reporter: Lars Hofhansl
> Fix For: 4.8.0
>
> Attachments: 3005-v2.txt, 3005-wip-v1.txt, PHOENIX-3005_v1.patch
>
>
> Currently the optimization in PHOENIX-258 is not used for DISTINCT index
> scans. We should add that as well.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
