on 02-16-2011 2:11 PM
Hi Experts,
We are continuously getting requirements from client for reports using the change log header and item table (CDHDR and CDPOS) for MM and ICL_HISTPOS_GUI for insurance.
I have noticed that these tables contain huge amount of records and which is really difficult to deal with.
Can anyone please suggest a better and efficient modeling to capture the records from these table on personal experience.
Right now I have planned for
u2022For a FM custom extractor to do most of the selection of records in the code/ or may be a view extractor can work with proper selection condtions.
u2022Using the write optimized DSO for staging in BW
u2022Then pulling the Delta into the cube
But what worrying me is the volume of records thatu2019s flowing in per day even per hour. Per day the number of records reaches sometime 4-5ks for couple of weeks it goes to 20-30ks. How to handle these huge amount of records. How to do the archiving for the same? How to do the performance optimization for the same.
Its done.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi,
The number of records should not be a problem in your case but the only thing is that these two system tables are set to be refreshed after a certain amount of time, so check the details from your basis/ECC team. Also, try to extract the minimum amount of data by providing proper filter conditions in order to fasten up the process as these tables contain all the changed values in the system which is huge.
Regards,
Arminder
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
try to delete delete the changelog use ST14 tcode...
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
95 | |
11 | |
11 | |
10 | |
9 | |
8 | |
6 | |
5 | |
4 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.