Addition of new field

[Former Member]

Hi ,

I have to include a new field into my cube. And i understand that for doing this, i have include the same in my infosource and inturn modify my datasource as r/3.

Now my doubt is:

1) Do i have to delete the data in the cube before adding the new field and then reload?

2) When i change my extract structure i.e after adding the new field, what is the user exit which i have to write to populate the value in that field? I mean the code.(if my datasource is 2LIS_04_P_MATNR and field is 0BATCH)

3)Also a quick question. I include a new field to my standard extract structure,by doing so will it affect any other programs which are using this structure?

Thanks in advance and I appreciate all ur help.

I have seen many threads and many suggestions by Roberto ,Anil and Kristian.

Please as well give me ur suggestions.

Regards,

Priya.

[edwin_harpino · Accepted solution]

hi Priya,

1) mostly you won't need to delete the data, unless you have to change the dimension of cube with the new field, if it's keyfigure then no need to delete data. also depend on whether you need the new field in your previous data or not, if you need it then have to reload, to save storage space you may want delete old data first.

2) the user exit is ZXRSAU01, check the detail steps from threads you have seen, code sample for you case can be like following :

data : l_s_2LIS_04_P_MATNR like [structure of 2lis_04_p_matnr],

l_tabix like sy-tabix.

case i_datasource.

when '2LIS_04_P_MATNR'.

loop at c_t_data into l_s_2LIS_04_P_MATNR.

L_TABIX = SY-TABIX.

• logic to get bacth, e.g get l_batch = ...

l_s_2LIS_04_P_MATNR-zzbatch = l_batch.

MODIFY C_T_DATA FROM l_s_2LIS_04_P_MATNR INDEX L_TABIX.

endloop.

when others.

endcase.

3) you add the 0BATCH field in an append structure mechanism so mostly it won't impact to any other programs, it's safe.

[Former Member · Accepted solution]

Hi,

Thank you. Now the field which i have to include is a characteristics and i think i then i will have delete the data and then do the reload of the same.

Another doubt. If there are some 5 cubes which gets its data filled using the same datasource and if i just include a new field in to 2 cubes (or have to just change 2 cubes) and change my data source. Now will the change in datasource affect my cubes for which the field is not added?

I will check out with the logic and will update u with the same.

Thanks and will reward once everything is clear.

Regards,

Priya.

[Former Member · Accepted solution]

Hi Priya,

your doubt can be dispelled with no problem:

empty tthe queue,

enhance your datasource/extract structure (and adjust the settings in RSA6),

write the userexit,

replicate in BW,

add the field in the transfer structure,

map your new field with an infoobject in the infosource/transfer rules,

empty your cubes,

add the new char (and related dimension),

reactivate all the update rules (even for all the other cubes!),

reload what you want and...

that's all!

Remember that:

*you cannot reconstruct (is it right, Dinesh?), since you need to load again from the source system and not simply from PSA

*you need to close your production system before to import the enhancement, empty the queue and then re-run the setup job and init your flow again!

*you can empty only two cubes (and init with no data transfer and load a full only in these two cubes) or empty everything and load again all

Hope it helps!

Bye,

Roberto

[Former Member · Accepted solution]

Hi Roberto Negro,

I hope, Reconstruction is possible until we do not make changes to Transfer structure.

With rgds,

Anil Kumar Sharma .P

[Former Member · Accepted solution]

Clearly, mine was a rhetorical question to Dinesh!

Anyway, PSA data doesn't contain the new info (coming from the exit in R/3 and related new setup job/init).

So, in this case, reconstruction is not possible.

Bye,

Roberto

[Former Member · Accepted solution]

Hi Roberto,

Thank you so much.

Regards,

Priya

[Former Member]

Hi Priya,

Yes you would need to delete th contents and reconstruct the requests in your case.

As for the cubes which do not have the fields - In the case you would not map those fields in the update rules. It would be set to "No update".

Hope this helps.

Bye

Dinesh