[
https://issues.apache.org/jira/browse/HIVE-24147?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17194002#comment-17194002
]
Jesus Camacho Rodriguez commented on HIVE-24147:
------------------------------------------------
The patch I put together in https://github.com/apache/hive/pull/1486 is
basically a bandage.
Instead, we should probably return proper column names in the JDBC result set
metadata call. It seems the appending is controlled by
{{hive.resultset.use.unique.column.names}}, which is set to {{true}} by
default. I am not sure about the implications of setting it to {{false}}, but
an option to uniquify those column names would be to use numeral after the
column name instead of prepending the {{table alias + '.'}}. In addition, we
could rely on the {{ResultSetMetaData.getTableName}} method to return the table
name if desired. However, if any application is relying somehow on the column
names having a certain format, this may result in a change of behavior for
those.
Thus, maybe the bandage works for the time being.
> Table column names are not extracted correctly in Hive JDBC storage handler
> ---------------------------------------------------------------------------
>
> Key: HIVE-24147
> URL: https://issues.apache.org/jira/browse/HIVE-24147
> Project: Hive
> Issue Type: Bug
> Components: JDBC storage handler
> Affects Versions: 4.0.0
> Reporter: Jesus Camacho Rodriguez
> Priority: Major
> Labels: pull-request-available
> Time Spent: 10m
> Remaining Estimate: 0h
>
> It seems the `ResultSetMetaData` extracted from the query to retrieve the
> table columns names contains these columns as fully qualified names instead
> of possibly using the {{getTableName}} method. This ends up throwing the
> storage handler off and leading to exceptions, both in CBO path and non-CBO
> path.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
