getting more data in bw while delta extraction than data records in r/3

[Former Member]

Hello all,

I am using 0FI_GL_10 datasource and have enhanced it with 0FUND ..the data is getting loaded from r/3 to ods to cube. I did an init yesterday and the the data records coming over were perfect 110313...today i loaded for delta and there were 51 records added to rsa3..but bw pulled 391 records...

How can i check why are the records pulled over more than 51?

Please give some suggestions.

Thanks,

Raj

[Former Member · Accepted solution]

Hi Raj,

Check the no. of records in RSA3 and then compare it with the no. of records in PSA. They should be the same. It could be that you are using some return tables in your update rules for the cube/ODS and hence the increase in the no. of records.

Bye

Dinesh

[Former Member · Accepted solution]

Thanks both,

When i load the data as init, the number of records are perfectly same in rsa3 and psa. Now when i load the delat load next day the number of records in monitor screen or psa for that load are way too much than actual the difference of records that are added in r/3 side (rsa3).

Now if i delete all the requets and reload the data (init) then everything is just fine.

Why is my extractor pulling more records only during delta? if i see more records in PSA itself than i guess ods/cube are way out of question is it ?

any more suggestions?

Thanks

[Former Member · Accepted solution]

Hi Raj,

Looks like something wrong with the logic written - probably in the user exit where you added 0FUND. Just put a breakpoint there and debug.

Bye

Dinesh

[Former Member · Accepted solution]

Hi dinesh,

If the logic is wrong shoudnlt it pull more records while init as well? I just checked my user exit, 0FUND is a field in FAGLFLEXT table itself (this is the same table from where other fields in datasource are coming from) so i did not have to write any user exit for 0FUND. I just added the field and it started populating.

Now i am confused is that even the right way of doing it?

Please suggest.

Thanks.

[Former Member · Accepted solution]

Hi Dinesh

51 is it the transfered record or the added record.

go to cube manage and check records in transfered column and added column.

Chris

[Former Member · Accepted solution]

Chris,

In ods transferred records are 391 and added are 255.

Thanks

[Former Member · Accepted solution]

if your rsa3 is extracting 391 records and BW is also extracting 391 from the source and adding 255 records then it should be fine.

correct me i am missing a point here.

Chris

[Former Member · Accepted solution]

Chris,

When i did init records in rsa3 and psa were 110313 and that good...today i planned to run delta and records in rsa3 are 110364..so that means 51 new records...but my extractor now pulls 391 records in psa and then when i load the data all the way to the cube i can see a lot of discrepancy with the data..so its definately not right...

Thanks for replying

[Former Member · Accepted solution]

Hello all,

"just checked my user exit, 0FUND is a field in FAGLFLEXT table itself (this is the same table from where other fields in datasource are coming from) so i did not have to write any user exit for 0FUND. I just added the field and it started populating.

Now i am confused is that even the right way of doing it?"

Can some one clear this point please?

Thanks,

[Former Member · Accepted solution]

> Hello all,

> "just checked my user exit, 0FUND is a field in

> FAGLFLEXT table itself (this is the same table from

> where other fields in datasource are coming from) so

> i did not have to write any user exit for 0FUND. I

> just added the field and it started populating.

> Now i am confused is that even the right way of doing

> it?"

>

> Can some one clear this point please?

>

> Thanks,

It is a kind of loophole. The extractor either uses a Move-corresponding to populate the extract structure or analyzes the structure and fills the corresponding fields. So it works here.

Dirk

[Former Member · Accepted solution]

> Chris,

> When i did init records in rsa3 and psa were 110313

> and that good...today i planned to run delta and

> records in rsa3 are 110364..so that means 51 new

> records...but my extractor now pulls 391 records in

> psa and then when i load the data all the way to the

> cube i can see a lot of discrepancy with the data..so

> its definately not right...

>

> Thanks for replying

First of all the extractor probably uses a date as a time stamp so if you run the delta, everything from the last day is transferred, even the records that are already in the init. You could try the following:

Run an init, then run a delta immediately afterwards. Make sure there are no postings in between. Delete the delta. Run the delta next day again and check the results.

Best regards

Dirk

[Former Member · Accepted solution]

Hi Dirk,

Thanks for the reply...so if you have read the whole post does the addition of 0FUND without user exit would that be an affecting factor for more records I am getting in my delta extraction?

When you mean loop hole so I can live with it or should i actually write an user exit for that ?

Thanks,

[Former Member · Accepted solution]

Hi Dirk,

As you said I did that, i ran an init and everything was good...and then i ran delta right away and I had 0 records of 0 and then i deleted than delta and then i see that users had posted some entries (68 records to be more precise) and then i ran my delta again and agaian the same thing i get 221 records rather than 68.

Thanks,

[Former Member · Accepted solution]

Does the extractor have currency type in it?

Whats probaly happening is it going from column sbased to row based (ie disaggregating on load)

In FI we tend to have columns for local and document currency

In BW we use two rows for that one record and differentiate by a partionign characterstic

[Former Member · Accepted solution]

Hi Simon,

you are right, there is currency type and my init and delta infopackge has the restriction 10 while loading the data (and when i am checking the records in rsa3 i am puttin 10 as condition in rsa3).

Can you explain a little bit more what you exactly you are doing because I am guessing that is some what issue I might be having.

Thanks

Edited message:

one more thing i would like to ask you is, are you having the keyfigures 0CREDIT and 0DEBIT as summation or overwrite?

Thanks again

Message was edited by:

Raj M

[Former Member · Accepted solution]

Hello all,

I was intially working n Qa environment and to debug this issue more i logged in dev and started working there. I loaded the data from scratch init, and everything went good till the cube. (data from r/3 and bw matches good). now i created some records on r/3 side...34 to be more precise and then i ran my delta and those 34 records always keep on coming in my delta (I am not loading that delta load from psa to cube yet..is that the reason )...so basically i ran delta infopackge like 3-4 times and everytime it comes with 34 records?

why is that?

thanks alot

[Former Member · Accepted solution]

> Hi Dirk,

> Thanks for the reply...so if you have read the whole

> post does the addition of 0FUND without user exit

> would that be an affecting factor for more records I

> am getting in my delta extraction?

>

> When you mean loop hole so I can live with it or

> should i actually write an user exit for that ?

>

> Thanks,

You shouldn't write a user exit for this. But you should be aware of two possible problems.

1. You could get problems with the delta handling when you only change the 0FUND. If it's possible try this and check if the delta comes to BW.

2. When you upgrade the Plug-In check this extractor carefully to see if the 0FUND field is still populated.

Best regards

Dirk

[Former Member]

hi

See number of records in RSA3 extractor checker

Then compare the number of records in the monitor

They both should be same

If they both are same but u get more records in cube/ods then your kfs get split up and get calculated into different kfs in the update rules

Assign points if useful

Regards

N Ganesh