[
https://issues.apache.org/jira/browse/HIVE-9674?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Mithun Radhakrishnan updated HIVE-9674:
---------------------------------------
Attachment: HIVE-9674.2.patch
Here's the fixed patch. I've changed HiveMetaStore to send a single
DropPartitionEvent.
Also, please note that by using
{{AuthorizationPreEventListener.PartitionWrapper}}'s first constructor, we
avoid multiple calls to {{HMSHandler.get_table_core()}}. This really adds up.
My perf tests indicate huge savings here.
We should consider a similar change in HIVE-9609.
> *DropPartitionEvent should handle partition-sets.
> -------------------------------------------------
>
> Key: HIVE-9674
> URL: https://issues.apache.org/jira/browse/HIVE-9674
> Project: Hive
> Issue Type: Bug
> Components: Metastore
> Affects Versions: 0.14.0
> Reporter: Mithun Radhakrishnan
> Assignee: Mithun Radhakrishnan
> Attachments: HIVE-9674.2.patch
>
>
> Dropping a set of N partitions from a table currently results in N
> DropPartitionEvents (and N PreDropPartitionEvents) being fired serially. This
> is wasteful, especially so for large N. It also makes it impossible to even
> try to run authorization-checks on all partitions in a batch.
> Taking the cue from HIVE-9609, we should compose an {{Iterable<Partition>}}
> in the event, and expose them via an {{Iterator}}.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
