[
https://issues.apache.org/jira/browse/HIVE-13257?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15400173#comment-15400173
]
stephen sprague commented on HIVE-13257:
----------------------------------------
use "order by 1" as your test case. you'll get the same error.
{code}
[dwrdevnn1:10001] hive> select count(*), date_key FROM
Raw_logs.fact_tracking_datekey_by_hour group by date_key order by 1 desc;
Fri Jul 29 15:47:06 2016 - Thrift::API::HiveClient2::execute: Error while
compiling statement: FAILED: SemanticException [Error 10128]: Line 1:7 Not yet
supported place for UDAF 'count'; HQL was: "select count(*), date_key FROM
Raw_logs.fact_tracking_datekey_by_hour group by date_key order by 1 desc"
{code}
and yes "hive.groupby.orderby.position.alias=true"
thoughts?
> GroupBy with column alias does not support AVG
> ----------------------------------------------
>
> Key: HIVE-13257
> URL: https://issues.apache.org/jira/browse/HIVE-13257
> Project: Hive
> Issue Type: Bug
> Affects Versions: 2.1.0
> Reporter: Prasanth Jayachandran
>
> For the following query, with hive.groupby.orderby.position.alias set to true
> {code:title=Query}
> SELECT Avg(`t0`.`x_measure__0`) AS `avg_calculation_270497503505567749_ok`
> FROM (SELECT `store_sales`.`ss_ticket_number` AS `ss_ticket_number`,
> Sum(`store_sales`.`ss_net_paid`) AS `x_measure__0`
> FROM `store_sales` `store_sales`
> JOIN `item` `item`
> ON ( `store_sales`.`ss_item_sk` = `item`.`i_item_sk` )
> GROUP BY `store_sales`.`ss_ticket_number`) `t0`
> GROUP BY 1
> HAVING ( Count(1) > 0 );
> {code}
> it throws the following exception
> {code:title=Exception}
> FAILED: SemanticException [Error 10128]: Line 2:7 Not yet supported place for
> UDAF 'Avg’
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
