[
https://issues.apache.org/jira/browse/TRAFODION-2166?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15425623#comment-15425623
]
David Wayne Birdsall commented on TRAFODION-2166:
-------------------------------------------------
The following script reproduces the issue on a workstation using
"install_local_hadoop":
-- to turn on ability to use ON NECESSARY COLUMNS
CQD USTAT_AUTOMATION_INTERVAL '100';
-- so UPDATE STATS doesn't take so damned long
cqd hive_max_string_length '200';
-- causes the trafodion.”_hivestats_”.sb_* histogram tables to be created
update statistics for table hive.hive.date_dim on every column;
-- create an objectUID for an interesting table
create external table customer_demographics for hive.hive.customer_demographics;
-- should get 6011 error on join column, but no reason = ' ' row gets generated
prepare s1 from
select count(*)
from hive.hive.customer_demographics x join hive.hive.customer_demographics y
on x.cd_purchase_estimate = y.cd_purchase_estimate;
-- tells us there's nothing to be updated (which is wrong)
update statistics for table hive.hive.date_dim on necessary columns;
> UPDATE STATS ON NECESSARY COLUMNS does not work on Hive tables
> --------------------------------------------------------------
>
> Key: TRAFODION-2166
> URL: https://issues.apache.org/jira/browse/TRAFODION-2166
> Project: Apache Trafodion
> Issue Type: Bug
> Components: sql-cmp
> Affects Versions: 2.1-incubating
> Environment: All
> Reporter: David Wayne Birdsall
> Assignee: David Wayne Birdsall
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
