cancel
Showing results for 
Search instead for 
Did you mean: 

Query Performance

Former Member
0 Kudos

Hi,

We have a cube which contains tons of data. To improve the performance we created a new multiprovider and now we are repalcing the old query(A) with new query(B) which is built on the multiprovider.

Now my question is when we are running the query we are seeing total time of B is more than the total time of A.

the query is using virtual keyfigures. We are running compression on the new cube in multiprovider.

Does it make any diff? Last week it was ok but this week this behaving diffrently..

The new query(B) is taking more time on OLAP than Old query(B)?

Why it is behaving so?

Any help will be really appreciated.

thanks

Accepted Solutions (0)

Answers (4)

Answers (4)

Former Member
0 Kudos

Solved

Former Member
0 Kudos

Hi Sri pal,

Can u plz describe how you could solve your problem.

/Veena

Naga_P
Active Participant
0 Kudos

Hi,

I assume you have not added any other Infoprovider (like an infoobject or ODS object) into the Multiprovider, apart from the base cube on which you have the original query. If so, there is a possibility the system is trying to merge the records and we end up with large data selectiion - evenif the query selections remain the same.

If you have not added any other infoprovider into the Multiprovider, this might happen if you have reorganized the characteristics / dimensions on the multiprovider. I mean to say if the structure of the Multiprovider is different from the underlying cube

rgds

Naga

Former Member
0 Kudos

We have divided the base cube into 4 cubes . we have partioned the cube based on year selection.

The query I am running is on current year where as the old cube contains all the 4 years of data and the new cube contains 2008 data.

Former Member
0 Kudos

We have divided the base cube into 4 cubes . we have partioned the cube based on fiscal period

The query I am running is on current year where as the old cube contains all the 4 years of data and the new cube contains 2008 data.

Former Member
0 Kudos

Hi,

Goto the following link and try to implement the same. The OLAP cache tuning will give u better performance.

https://www.sdn.sap.com/irj/sdn/go/portal/prtroot/docs/library/uuid/9f4a452b-0301-0010-8ca6-ef25a095...

Regds,

Shashank

Former Member
0 Kudos

use the process chain to automatize the load; include processed sequence wise, and for performance tuning=> include a process for the ABAP code RSR_GEN_DIRECT_ALL_QUERIES at the end of the cube load

This will support your Cache concept and improve the performance tremendously.

As one time query is executed for a given range of dates, next time for the same date interval , output will be read through cache.

former_member251078
Participant
0 Kudos

Hello Sri Pal,

As you mentioned you have created a multiprovider which will be accessing the patitioned infocubes. As per your analysis, the DB time is higher than the OLAP time. Have you checked in RSDDSTAT_DM table, whether an aggregate is accessed during the execution of the query ??

I would recommend you to check in RSDDSTAT_DM table first to check whether an aggregate is accessed during the execution of the query, if not you have to create an aggregate on the corresponding basic cube to improve the performance of the new query. You can check the aggregate suggestion in RSRT transaction.

Yours Sincerely

Dileep

Former Member
0 Kudos

Hi,

We dont have aggregates built on that cubes. We are not planning to built aggregates because the size of data is not big in the partioned cube.

Also If I see the no of selected/transferred in the old query the ratio is really big when you compare with the new one.

It is no making sende for me. Since the query is using virtual keyfigures, both the querys are using same code.

Where else can I concentrate? Please help me?

Former Member
0 Kudos

Just by Creating a MultiProvider on top of Cube does not increase performance.

What did u do in MultiProvider?

Former Member
0 Kudos

Hi,

We have implemented partioning based on fiscal year.

We have infoprov variable in our queries which is restricted by a variable to go particular cube.

The problem is last week the olap time is low comaprative this week. This week the old query is taking less time when you comapre with new query

There are no changes in system. we have implemented support packs to our system. Does it make any diff?

Please let me know?

Former Member
0 Kudos

Sorry the DBtime is more not the OLAP Time