Performance for derivation of 10 characteristics using ABAP routine

[Former Member]

Hi,

I've around 10 characteristics that need to be populated using ABAP routine based on 11th characteristic in the infocube.

Could someone suggest what is the best approach interms of performance to do this.

1. is it better to write transfer rotune for each characteristic (total of 10 characteristics)

2. is it better to write start routine? if Yes. where is it advisable? transfer rules or update rules?

FYI : procedure for Derivation of 10 characteristics is similar and uses 1 custom function module.

1st characteristic calls this function module 1 time.

2nd characteristic calls this function module 2 times.

3rd characteristic calls this function module 3 times in a loop etc.

Appreciate any feedback

Hari Immadi

http://immadi.com

SEM BW Analyst

[former_member205352 · Accepted solution]

Better approach would be to handle this in Start routine/End routine(if on BW 7).

what does function module do ?? Any Selects ?

[Former Member · Accepted solution]

thanks Praveen.

We are still on BW 3.x.

so, as per your suggestion, start routine is the best bet.

any recommendation on where to use this? transfer rules or update rules.?

Hari Immadi

http://immadi.com

SEM BW Analyst

[former_member205352 · Accepted solution]

It depends,

if you want same logic to be applied when you load to multiple datatarget then go for transfer rules start routine.

If its specific to single datatarget then go for update rules.

[virenp_devi · Accepted solution]

Hi,

It depends, if you have multiple data targets and you have to write the code to multiple update rules I suggest to write in Transfer routine. This will save to write same code multiple times and for better performance also but if it just a one data target please write in update rule. As if you write anything in transfer routine we always have to push data through PSA.

Hope this helps. Let me know if you have any doubt.

Regards,

Viren

[Former Member]

Thanks all,

that was very good information.

Hari Immadi

http://immadi.com

SEM BW Analyst

[Former Member]

Hi Hari,

The best way to do this is Start Routine.

When you are loading data in the cube, if you are loading data from a datasource or a DSO, add all the 10 characteristic to the DSO or InfoSource. Doing this will get all the characteristics in Data_pckage in start routine.

The in start routine, run the Function module and get these characteristics poppulated in the Start routine. Map these 1:1 then in the update rules.

Hope this helps, please let me know if you need nay help.

Regards,

Pankaj