cancel
Showing results for 
Search instead for 
Did you mean: 

line items data extraction taking more than a day for just 1 period

former_member199935
Participant
0 Kudos

Hi,

we are doing a fresh implementation of NW 7.3 and we have a requirement to extract PCA line items from R3 (PCA3 datasource).

while extracting data into infocube, it is taking more than a day just for 1 period containing 35 million records. so, if we go this way looks like, it is going to take few weeks to load all of historical periods. How do we improve this data load time.

appreciate any feedback.

Accepted Solutions (0)

Answers (1)

Answers (1)

Former Member
0 Kudos

how much data does the source system have? Monitor activity SM50 and see which tables its spending so much time. If you need to archive some old data; please do...or try some additional indexes based on SQL statements....

former_member199935
Participant
0 Kudos

for a single period, we've 35 million records.

I should have mentioned earlier that, R3 to PSA extraction completed in 6 hours for 35 million records.

PSA to InfoCube is taking more than a day. so, are you suggesting to create additional indexes on R3 tables? I doubt if this helps as we've data already in BW and the issue is within BW.

Former Member
0 Kudos

Ohk, i thought R/3 to PSA extraction is taking like a day. If its taking longer to load in infocube, you may have to redesign the individual dimensions. Also, how many parallel processes do you have? Try tweaking packet sizes and monitor where time is being spent in SM50.......6 hours for 35 million from ECC is quite ok....Also, you can always run infopackages with different selection conditions (e.g. company codes/ sales org , etc) in parallel. That will take some time away...