[
https://issues.apache.org/jira/browse/HIVE-12213?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14970300#comment-14970300
]
Aleksei Statkevich commented on HIVE-12213:
-------------------------------------------
I think it's more of a convention whether to write zero-value stats or not.
Currently they are written, so let's keep it this way.
However, there's slight inconsistency in this rule because of !newDir check on
line 229 in MetaStoreUtils. It means that if the directory was just created
(so, it's empty) then we'll not update table stats with zero values. I'd like
to keep it this way, but it's something that might need rethinking in future.
So, I'll use an earlier version of the patch which doesn't write table stats
for partitioned tables, but keeps everything else as it is.
> Investigating the test failure TestHCatClient.testTableSchemaPropagation
> ------------------------------------------------------------------------
>
> Key: HIVE-12213
> URL: https://issues.apache.org/jira/browse/HIVE-12213
> Project: Hive
> Issue Type: Test
> Components: Test
> Affects Versions: 2.0.0
> Reporter: Aihua Xu
> Assignee: Aleksei Statkevich
> Priority: Minor
> Attachments: HIVE-12213.2.patch, HIVE-12213.patch, HIVE-12231.1.patch
>
>
> The test has been failing for some time with following error.
> {noformat}
> Error Message
> Table after deserialization should have been identical to sourceTable.
> expected:<[TABLE_PROPERTIES]> but was:<[]>
> Stacktrace
> java.lang.AssertionError: Table after deserialization should have been
> identical to sourceTable. expected:<[TABLE_PROPERTIES]> but was:<[]>
> at org.junit.Assert.fail(Assert.java:88)
> at org.junit.Assert.failNotEquals(Assert.java:743)
> at org.junit.Assert.assertEquals(Assert.java:118)
> at
> org.apache.hive.hcatalog.api.TestHCatClient.testTableSchemaPropagation(TestHCatClient.java:1065)
> {noformat}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
