on 2011 Mar 02 6:57 PM
Hi gurus,
We are extracting data from some SQL databases which include several text fields. Due to the BW limitation of 60 char for characteristic objects, these fields are truncated when loaded. We have looked at 'chunking' up the text and storing it into several infoobjects as an option.
We have another process that extracts data out of BW and sends it to another application. This application requires the full text field for users to work with.
Since we have the full text field available in the PSA table before storing it into the several BW infoobjects, our thinking is that if there is a way to read the PSA table and store it into an ABAP table, the interface program reading from BW could use that to simplify the text field extraction. Our concern is that we understand that the PSA table is dynamically generated by BW, so there is no way to guarantee we will always read from the correct table.
Any guidance, comments or ideas are appreciated!
Thanks,
Darryl
Hi Darryl,
PSA table is fixed for a system (like Development/Quality/Production) unless you delete the DS. Get the PSA table names from the required system and access it through a program and write it to a table. That is possible. To get the PSA table name,
a. Double click on the data source
b.From the menu bar, select u201CGoto -> Technical Attributesu201D.
c. Under PSA Properties, you can find the PSA Technical Name.
Hope this helps.
Thanks and Regards
Subray Hegde
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Thanks Subrey,
That's what I was thinking. However, since our QA and and PRD systems are locked, we would have to create programs (one each for Dev, QA, and PRD?) in our Dev system with fixed table names and transport the matching one to QA / PRD.
The only danger I see would be if the PSA table changed in the future, since it would essentially be 'hard coded' into the ABAP program. I would prefer if there was a way for the PSA table to be referenced in a way that would allow it to always use the correct name no matter which system or what changes may occur that would generate a new table name. Any ideas on how that could be looked up that would work across all systems?
Thanks though for the info!
Cheers,
Darryl
hi Darryl
you can achieve that by looking at table rstsods
in the object name you will find the datasource name with the related source system
so you can do a search with a contain pattern the name of the datasource that will certainly not change.
then after you give in the name of the DS in the object name take the technical ODS name (in first versions of BW a psa was known as dso) the value corresponds to the psa table you look for.
take care of the validity date. the right one should always be the one with to date 01.01.9999 here no danger on any of the systems you can check, you will always pick up the right psa.
take care as well to the status ACT
the issues I see in copying the psa table will be the number of records and how often it is deleted regarding the frequence of execution of your program. good luck
hope this could help you out
Boujema
User | Count |
---|---|
70 | |
10 | |
10 | |
10 | |
10 | |
8 | |
8 | |
7 | |
5 | |
5 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.