Performance issue in query ot get row of ACT_HI_VARINST

[hsimon]

Hi,

I just registered to this forum since I have to support the activity engine by a customer. We found that the cost for following query is very high:

select * from ( select a.*, ROWNUM rnum from ( select RES.* from ACT_HI_VARINST RES WHERE RES.TASK_ID_ = ? order by RES.ID_ asc ) a where ROWNUM < ?) where rnum >= ?

The reason is the full table scan used by the oracle rdbms.

For tests, we created a temporary index like CREATE INDEX IDX_TASK_ID ON ACT_HI_VARINST (TASK_ID_) and the query path is now a range scan with much lesser cost.

Does anyone have the same trouble? I guess, the index should be created autoamtically while installing.

Regards

Harald

[trademak]

Hi Harald,

Could you create a JIRA issue so we can track this issue? I'll add the index to the create scripts already.

Thanks,

[hsimon]

Hi Tijs,

a new issue has been created.

thanks!

Regards

Harald