[
https://issues.apache.org/jira/browse/HIVE-1370?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12871455#action_12871455
]
Carl Steinbach commented on HIVE-1370:
--------------------------------------
You need to use Hive's LATERAL VIEW syntax. See
http://wiki.apache.org/hadoop/Hive/LanguageManual/LateralView
I agree that the error message is misleading. Perhaps this ticket would be a
good place to fix it.
> Should be possible to select a primitive column together with UDTF
> ------------------------------------------------------------------
>
> Key: HIVE-1370
> URL: https://issues.apache.org/jira/browse/HIVE-1370
> Project: Hadoop Hive
> Issue Type: Improvement
> Reporter: Alex Kozlov
> Priority: Minor
>
> I have a table that contains a primitive type column and an array. For
> example 'create table temp (guid string, events array<struct<num:int,
> ts:bigint>>)'. I am able to select guid and UDTF separately:
> hive> select guid from temp;
> and
> hive> select explode(events) as event from temp;
> What is the reason I can not do:
> hive> select guid, explode(events) as event from temp;
> FAILED: Error in semantic analysis: UDTF's are not supported outside the
> SELECT clause, nor nested in expressions
> All other implementations of PIVOT or TRANSPOSE I know do allow that. The
> error message is very misleading in any case: I am not using UDTF outside of
> select or in an expression.
> Thanks,
> Alex K
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
