In my Bex report i have Six characters among them which one i can go for aggregates . As i know their are some limitations for aggregates we cannot have too many aggregates which will degrade the performs. according to the
Thumb rule = Number of records read from DB / number of records transferred
,Aggregation ratio > 10 %,
Percentage of DB time > 30%,
where can I check exactly which characteristic has grater than given percentage for each field Apart from ST03,
Can one help me the Tips that can be taken before going for Aggregates .
System itself will help you to propose the aggregates, check the below link to see how....
Thanks Durgesh for responding the mail,
I would like to know how does system propose which characteristics should we go for aggregates. Do we have any tool which captures statistical results of DB time and Thumb rule . If you saw my mail i have equal number of data for all six characteristic that has to be aggregated in my report. I would realy appreciate if you give me with soem Tcode or any other links by which I can track each field.
U can find the DB time, OLAP time etcc. with the Tcode : ST03N (analysis Tool) . It all depends on the report requirement . If u have a report that users run daily and its performance is worst , then u have to improve the performance one such tool is u can use the aggregate.
Hope u got it,
Thanx & Regards,
Thank Ravi chandra,
As you mentioned ST03N we can find Overall % DB time , % OLAP ,% Front end for specific a cube According to Thumb rule if % DB time is grater than 30 % we can go for aggregates . But my question is after identifying % DB time > 30 I have six fields in my report among this six fields which one I have to go for aggregates. Correct me if I am wrong we get this information if we can have individual field level information in any T code.
Make sure your BW Statistics are running properly/regularly.
Right click to cube, choose Maintain Aggregate. The Proposals for Aggregates dialog box appears, and select propose aggregates.
Once the aggregate is created, you can monitor to see if it's being used, if not optimized you can delete it easily. Make sure you add a rollup step to the process chain.
I found that this is the easiest way, instead of trying to "assume and spend time". You can also go to RSRT I believe and run a specific query in test mode and ask there to propose an aggregate if a specific query is very important/concerned.
In order to get suggestions for best aggregates from the system, you need to run the query in RSRT,
the setting's should be as given below:
1) Give the query name
2)Check " Display statistics of the database table"(It displays RSDDSTST for this execution)
3)check "Displays aggregate found"(It displays best possible aggregates for this query execution)
Then choose execute+Debug to get a dialog box, and you can get many options.
RSRTRACE gives aggregate suggestions as well, apart from that, you still need to take a decision of which
charecteristic you want to create aggregates. After all the tools we have their is nothing compares to human
decision making. Points to be considered are
1) Characteristic which is making a good subset.
2)Depending on your query requirement, choose the characteristic that will be frequently used.
3)Consider the one's which will be used for many queries.