on 2006 Oct 19 3:16 PM
A couple of questions:
I'm already running delta loads into an ODS. Can I still perform a full load? Will that affect the deltas? Where can I find more documentation on delta mechanism.
Also, I have a cube on top of the ODS. If I wan unload the ODS and reload, will I have to unload too?
Thanks in advance,
Frank
Hi Frank,
If you are laoding deltas to an ODS, you can load a full load but be sure to mark it as a repair request (in the InfoPackage CtrlShiftf1) else it will give an error in activation. This will not affect the ongoing deltas. If you have any key figures that are updated as Addition (as oppposed to the usual ODS overwrite) then this may duplicate the numbers.
If you delete the contents of the ODS, please delete the contents of the cube and reload this as well to avoid any data integrity issues.
Hope this helps...
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Frank,
You can try to estimate that in RSA3 (in your source system). Make sure you increase the Data Records / Calls and Display Extr. Calls, else you will not be able to get the correct number, eg. the std values are 100 and 10 respectively, so only 1000 records will be returnd. You want to make this a number greater than the expected number of records.
Also depending on the datasource, an estimate can also be based by looking at the number of table entries, or document numbers...
Hope this helps...
Hi Frank,
The datamart interface will take care of the changed records (delta). For a cleaner process it is good to reload the cube as well...what is the volume of data we are considering? I am sure the load to the cube is less intensive as compared to a full load to the ODS from the source...
Hi,
Yes you can load full loads, but you'll have activation problems. Also, in some cases the system won't let you load the delta saying that a full load already exists. If you want to load the full load then mark it as a repair request.
You need to dump the cube too if you plan to dump the ODS, as there is a chance that you might load duplicate data from the ODS to cube, etc. So to be on the safer side, better to dump the cube.
Cheers,
Kedar
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
62 | |
9 | |
8 | |
6 | |
6 | |
6 | |
6 | |
6 | |
5 | |
5 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.