New characteristic in infocube kills BEx Query performance

[Former Member]

Hi,

I created ZPA_C01 infocube (copy of standard 0PA_C01) and I created some BEx Queries on it. This cube receives data from two different data sources.Data loading were working ok. BEx Query performance was ok.

Then I added a new characteristic to infocube, update the transformation that populates it. Then, I deleted old requests of data where new characteristic was empty. I deleted infocube indexes, performed a data loading and regenerated infocube indexes.Data loading worked fine after having added characteristic.

I made a change in BEx Query making reference to the new characteristic but now performance is terrible awful (time response: 20 minutes). New characteristic is and indicator (possible values: X or empty) and I'm referencing it in a selection of a row.

Since, infocube receives info from two datasource, should I delete infocube whole content in order to reestructure indexes? 

ZPA_C01 is using cache memory and so does BEx Query.

Infocube contains a million of records. This is not a issue for BW. No aggregates in infocube because BEx Queries show detailed information. High cardinality characteristic (0EMPLOYEE) is a free characteristic, not part of rows.

Please let me know your suggestions.

Thanks in advance.

Regards.

Juana

[umashankar_poojar]

Hi Juana,

Did you try by removing new characteristics from query and check the performance?

Have you used this new characteristics in filter of the query and is there any exclusion on blank values?

Next try to generate DB Statistics after each load and check the performance of the query.

Thanks,

Umashankar

[Former Member]

hi Juana

check the statistics in rsrt to see where it is taking more time.

Since, infocube receives info from two datasource, should I delete infocube whole content in order to reestructure indexes?

you dont need to.