[
https://issues.apache.org/jira/browse/IMPALA-4364?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17185350#comment-17185350
]
ASF subversion and git services commented on IMPALA-4364:
---------------------------------------------------------
Commit 21c50f8dbb7ef75dbb56bee07590a710993f4d64 in impala's branch
refs/heads/master from Vihang Karajgaonkar
[ https://gitbox.apache.org/repos/asf?p=impala.git;h=21c50f8 ]
IMPALA-4364: [Addendum] Compare specific fields in StorageDescriptor
The query option REFRESH_UPDATED_HMS_PARTITIONS was introduced
earlier in IMPALA-4364 to detect changes in the partition
objects in HMS when a refresh table command is issued. Originally,
it relied on using the StorageDescriptor#equals() method to
determine if the Partition in catalogd is same as partition
in HMS with while executing the refresh statement.
However, using StorageDescriptor#equals() is dependent on HMS
version and may introduce inconsistent behaviors after upgrades.
For example, when we backported the original patch to older
distribution which uses Hive-2, the SkewedInfo field of
StorageDescriptor is not null. This field causes the comparison
logic to fail, since catalogd doesn't store the SkewedInfo
field in the cached StorageDescriptor to optimize memory usage.
This patch modifies the comparison logic to use explicit
implementation in HdfsPartition class which compares only
some fields which are cached in the HdfsPartition object.
Testing:
1. Added a new test for the comparison method.
2. Modified existing test for the query option.
Change-Id: I90c797060265f8f508d0b150e15da3d0f9961b9b
Reviewed-on: http://gerrit.cloudera.org:8080/16363
Tested-by: Impala Public Jenkins <[email protected]>
Reviewed-by: Vihang Karajgaonkar <[email protected]>
> REFRESH does not pick up ALTER TABLE...PARTITION...SET LOCATION changes
> -----------------------------------------------------------------------
>
> Key: IMPALA-4364
> URL: https://issues.apache.org/jira/browse/IMPALA-4364
> Project: IMPALA
> Issue Type: Bug
> Components: Catalog
> Affects Versions: Impala 2.6.0
> Reporter: Jacob Evan Beard
> Assignee: Vihang Karajgaonkar
> Priority: Major
> Labels: usability
> Fix For: Impala 4.0
>
>
> AFAIK the REFRESH command should pick up all changes to a table made by ALTER
> TABLE from outside of Impala (e.g. Spark SQL), however REFRESH does not pick
> up changes from ALTER TABLE...PARTITION...SET LOCATION, which seems to
> require an INVALIDATE METADATA instead.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
