You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@impala.apache.org by "Alexander Behm (JIRA)" <ji...@apache.org> on 2017/03/17 21:45:41 UTC
[jira] [Resolved] (IMPALA-1831) Query behaves strangely on the
Kerberos cluster (e1320)
[ https://issues.apache.org/jira/browse/IMPALA-1831?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexander Behm resolved IMPALA-1831.
------------------------------------
Resolution: Cannot Reproduce
This is on a very old version on a cluster that does not exist anymore. Very likely not relevant anymore.
> Query behaves strangely on the Kerberos cluster (e1320)
> -------------------------------------------------------
>
> Key: IMPALA-1831
> URL: https://issues.apache.org/jira/browse/IMPALA-1831
> Project: IMPALA
> Issue Type: Bug
> Components: Backend
> Affects Versions: Impala 2.2
> Reporter: Taras Bobrovytsky
> Assignee: Taras Bobrovytsky
> Priority: Minor
> Labels: performance, security
>
> The following query takes 2+ hours to execute on the Kerberos cluster (e1320) until there is a "Memory Limit Exceeded" error.
> Original Query:
> {code}
> select LEAST(COALESCE((supplier_no) * (supplier_no), -355), COALESCE(MIN((supplier_no) * (supplier_no)), 137)) AS int_col,
> COUNT(COALESCE((supplier_no) * (supplier_no), supplier_no)) AS int_col_2,
> (supplier_no) + (supplier_no) AS int_col_3,
> (supplier_no) * (supplier_no) AS int_col_4
> FROM revenue
> GROUP BY
> (supplier_no) + (supplier_no),
> (supplier_no) * (supplier_no)
> UNION
> SELECT
> (t1.l_suppkey) * (t1.l_quantity) AS decimal_col,
> t1.l_orderkey,
> LEAST(COALESCE(MIN(868) OVER (), -233), -755.787102873) AS float_col,
> COALESCE(SUM(t1.l_suppkey), t1.l_orderkey, LEAST(COALESCE(t1.l_orderkey, 650), COALESCE(t1.l_orderkey, 517))) AS int_col
> FROM lineitem t1
> WHERE
> (t1.l_suppkey) != (t1.l_tax)
> GROUP BY
> (t1.l_suppkey) * (t1.l_quantity),
> t1.l_orderkey
> {code}
> The query is basically a combination of the following 2 queries which are combined with the UNION clause.
> Query 1:
> {code}
> select LEAST(COALESCE((supplier_no) * (supplier_no), -355), COALESCE(MIN((supplier_no) * (supplier_no)), 137)) AS int_col,
> COUNT(COALESCE((supplier_no) * (supplier_no), supplier_no)) AS int_col_2,
> (supplier_no) + (supplier_no) AS int_col_3,
> (supplier_no) * (supplier_no) AS int_col_4
> FROM revenue
> GROUP BY
> (supplier_no) + (supplier_no),
> (supplier_no) * (supplier_no)
> {code}
> Query 2:
> {code}
> SELECT
> (t1.l_suppkey) * (t1.l_quantity) AS decimal_col,
> t1.l_orderkey,
> LEAST(COALESCE(MIN(868) OVER (), -233), -755.787102873) AS float_col,
> COALESCE(SUM(t1.l_suppkey), t1.l_orderkey, LEAST(COALESCE(t1.l_orderkey, 650), COALESCE(t1.l_orderkey, 517))) AS int_col
> FROM lineitem t1
> WHERE
> (t1.l_suppkey) != (t1.l_tax)
> GROUP BY
> (t1.l_suppkey) * (t1.l_quantity),
> t1.l_orderkey
> {code}
> If these queries are executed independently, Query 1 returns 0 rows in less than 1 second and Query 2 returns many rows after about 20 minutes.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)