[
https://issues.apache.org/jira/browse/HIVE-12369?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Matt McCline updated HIVE-12369:
--------------------------------
Description:
Implement Native Vector GroupBy using fast hash table technology developed for
Native Vector MapJoin, etc.
Patch is currently limited to a single key with a single COUNT aggregation.
Or, a single key and no aggregation also known as duplicate reduction.
3 new kinds of classes introduces that stored the count in the slot table and
don't allocate hash elements. Example:
{noformat}
COUNT(column) VectorGroupByHashLongKeySingleCountColumnOperator
COUNT(key) VectorGroupByHashLongKeySingleCountKeyOperator
COUNT(*) VectorGroupByHashLongKeySingleCountStarOperator
{noformat}
And the duplicate reduction operator a single key. Example:
{noformat}
VectorGroupByHashLongKeyDuplicateReductionOperator
{noformat}
was:
Implement Native Vector GroupBy using fast hash table technology developed for
Native Vector MapJoin, etc.
Patch is currently limited to a single Long key with a single COUNT
aggregation. Or, a single Long key and no aggregation also known as duplicate
reduction.
3 new classes introduces that stored the count in the slot table and don't
allocate hash elements:
{noformat}
COUNT(column) VectorGroupByHashLongKeyCountColumnOperator
COUNT(key) VectorGroupByHashLongKeyCountKeyOperator
COUNT(*) VectorGroupByHashLongKeyCountStarOperator
{noformat}
And the duplicate reduction operator a single Long key:
{noformat}
VectorGroupByHashLongKeyDuplicateReductionOperator
{noformat}
> Native Vector GroupBy (Part 1)
> ------------------------------
>
> Key: HIVE-12369
> URL: https://issues.apache.org/jira/browse/HIVE-12369
> Project: Hive
> Issue Type: Bug
> Components: Hive
> Reporter: Matt McCline
> Assignee: Matt McCline
> Priority: Critical
> Attachments: HIVE-12369.01.patch, HIVE-12369.02.patch,
> HIVE-12369.05.patch, HIVE-12369.06.patch, HIVE-12369.091.patch,
> HIVE-12369.094.patch, HIVE-12369.095.patch, HIVE-12369.096.patch
>
>
> Implement Native Vector GroupBy using fast hash table technology developed
> for Native Vector MapJoin, etc.
> Patch is currently limited to a single key with a single COUNT aggregation.
> Or, a single key and no aggregation also known as duplicate reduction.
> 3 new kinds of classes introduces that stored the count in the slot table and
> don't allocate hash elements. Example:
> {noformat}
> COUNT(column) VectorGroupByHashLongKeySingleCountColumnOperator
> COUNT(key) VectorGroupByHashLongKeySingleCountKeyOperator
> COUNT(*) VectorGroupByHashLongKeySingleCountStarOperator
> {noformat}
> And the duplicate reduction operator a single key. Example:
> {noformat}
> VectorGroupByHashLongKeyDuplicateReductionOperator
> {noformat}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
