[
https://issues.apache.org/jira/browse/TAJO-1410?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14375737#comment-14375737
]
Hyoungjun Kim commented on TAJO-1410:
-------------------------------------
I also think that we don't need TPC-DS full data set and full query set.
But I think that more query test cases can easily find a bug.
I know some TPC-DS queries are similar. So It is not necessary that all queries
to be tested.
If we can find query similarity, we can remove similar queries. But that is
time boring task.
Another reason why I made this test case is that it is difficult to make
manually proper data set.
If TPC-DS small data is supplied, this test can be replaced with general test
case of Tajo.
> TPC-DS TestCase
> ---------------
>
> Key: TAJO-1410
> URL: https://issues.apache.org/jira/browse/TAJO-1410
> Project: Tajo
> Issue Type: Test
> Reporter: Hyoungjun Kim
> Assignee: Hyoungjun Kim
> Priority: Minor
> Attachments: Tajo TPC-DS.pdf
>
>
> There is no TPC-DS TestCase in the current source code. It is difficult to
> make small TPC-DS dataset because TPC-DS is more complex than TPC-H. I
> propose TPC-DS TestCase as the followings:
> - The default build doesn't execute TPC-DS test case.
> - Add new maven profile 'tpcds-test' for TPC-DS test. If someone want to run
> TPC-DS test, run the next command.
> {noformat}
> mvn -Ptpcds-test -Dtpcds.gen.data=true
> -Dtpcds.generator=/tmp/tpcds-kit-master/tools/dsdgen
> -Dtpcds.data.dir=/tmp/tpcds install
> {noformat}
> - If tpcds.gen.data is true, The build script calls TPC-DS tool to generate
> test set with scale factor 1. TPC-DS tool should be installed in the test
> machine. Test data is located at tpcds.data.dir.
> - If test data is already exists in the test machine, set 'tpcds.gen.data'
> property false and set 'tpcds.data.dir' property with data directory.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
