on 08-11-2015 7:13 AM
Dear Gurus,
I've got a requirement to load data month by month for year where each month is loaded by package rather than not in one package.
for ex:
currently my DTP loading all the records for a year like 400000 records in one package or based on the package size.
It's like loading monthly data as 1 month in 1 package.
Kindly let me know how to write control by package and ABAP code for the same.
Thanks a million in advance.
Regards,
Ravi Kanth
Thank you one and all for your valuable inputs.
Now that I've changed my expert routine code to handle the data, month by month within the package.
Though it is pretty bad @ performance, but as the data size is less for a month, it is OK for the system to handle.
Thanks again.
Ravi
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
If you are sure about the data volume, the Other option is just bring all data in one package and process as required in Routines. Else Go with the Rule Group, as stated by Maunank.
Cheers!
Chandan
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hello Ravi,
This is understood from your requirement that you are doing something with the month and you need all the records of similler month in single package . ONce you get those month in single package you need to do some derivation on that month package.
I would suggest you go with the creation of new rule groups in transformation .
1. Create 12 rule group for twelve months
2. if rule groupe one is for January then add below statement in start routine
DELETE SOURCE_PACKAGE WHERE CALMONTH NE "001"
value for Calmonth changes according to your rule group
By this way only single month will be availabel in one rule group and you can do futher derivation accoding to your requirement.
Hope this will help.
Thanks & Regards,
Maunank Patel.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Check the data and based on the fileds like Year and Month , try to use the SYMANTIC GROUPING in DTP. As that is only viable option to such a unique requirement of yours'
Cheers!
Chandan
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Chandan,
Tried it, but not working!
This strange requirement comes with my Expert routine where I need to calculate consecutive days of employees for a month. So, here I need a expert routine which calculates month by month.
FYI... the link for my post on consecutive days....
http://scn.sap.com/thread/3778800
Thanks again
Regards,
Ravi Kanth
here's two options: either created 12 DTPs (on for each month) and make the year "variable" (either via a BEx variable or via code lookup) or create 1 DTP where both month & year are "variable"
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Raf,
this is a tedious job and much maintenance required.
It would be better load manually for time-being for historical.
This strange requirement comes with my Expert routine where I need to calculate consecutive days of employees for a month. So, here I need a expert routine which calculates month by month.
FYI... the link for my post on consecutive days....
http://scn.sap.com/thread/3778800
Thanks again
Regards,
Ravi Kanth
I don't see the link between "expert routine" and your other post?! Either you (try an) solve this on the reporting side (where you should use exception aggregation, probably nested) or you solve it in a transformation. In the latter part, it shouldn't really matter how many months are in each package.
Dear Naga,
Use Semantic group in DTP, and tick the month selection, this will group the data month by month in packages.
Thanks
Vijayendra
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi,
Further to Vijayendra's reply, I think that Semantic Grouping (on Calendar Month) is one part of the approach. It will avoid that data belonging to the same month is spread over multiple data packages.
However, depending on the package size multiple months can be stored in one data package. If I understood you correctly, this you want to avoid. So the other part is reducing the package size so that only one month will fit in one data package.
For my understanding and just because I am curious, why do want to control the loading process like this? If there is no requirement to influence packaging with semantic Grouping, I would advice to leave it out. It will degrade the loading performance.
Best regards,
Sander
Hi Sander,
Thanks for detailed help.
But I can't assume on the # of records for each month then I cannot restrict on package size.
If I opt semantic and package size a 100,000, system is get all the records for all then months into 1 package if records are less than 100K.
This strange requirement comes with my Expert routine where I need to calculate consecutive days of employees for a month. So, here I need a expert routine which calculates month by month.
FYI... the link for my post on consecutive days....
http://scn.sap.com/thread/3778800
Thanks again
Regards,
Ravi Kanth
Hi Ravi,
I had a look at your other thread. Even if you could ensure 1 month per package by some nasty voodoo, I don't see how that helps solve the problem of consecutive days being summed up in the reporting layer. Nothing you do in the expert routine or in load splitting will change the aggregation behavior of a query.
As far as calculations are concerned, a transformation with a Sem Grouping on Employee and month can do that with routines as required. However, the routines shouldn't require that only one month must come in a packet. If your expert routine works only with one month per packet, then it seems that the routine needs tweaking.
Post the routine here, maybe we can find some solution.
Regards,
Suhas
Hi Ravi,
As per requirement go to IP data selection tab specify the month ans schedule with start.
Goto DTP Extraction mode Delta with get all new data request by request, and execute DTP.
Regards
Sathya.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
I second Sander - it is possible using Sem Grouping but inadvisable. I too am curious about the kind of requirement that would necessitate such a data loading procedure.
It seems you are doing this just for splitting up a one-time load of five years. If that's the case there are simpler ways to do that.
Your requirement is not possible with code in the transformation.
User | Count |
---|---|
86 | |
10 | |
10 | |
9 | |
6 | |
6 | |
6 | |
5 | |
4 | |
3 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.