N:1 Maping

[Former Member]

Hi guys,

I have three sender files in which one field called ITEM is common in all the files and I have created a ZTable in the Target Structure according to the structure of the input files where in the Ztable there is one ITEM field and other fields based on the three files.please tell me how I can perform the N:1 Mapping .

Reply me at the earliest.

Thanks & Regards,

Vniay Yadav

[Former Member]

Hey

have a look at the following

/people/pooja.pandey/blog/2005/07/27/idocs-multiple-types-collection-in-bpm

in the above thread File_Type has been used as correlation field,and in ur case it will be ITEM

even though its for IDOC's but the logic will be the same for files too

thanx

ahmad

[Former Member]

Hi Vinay,

You can use BPM in this case... Create three different interfaces to read the 3 files and in the BPM use the ITEM field as the correlation .

There could be other solutions 2. If you tell us your requirement more clearly.

What all data are you collecting from the 3 files and then updating in Z Table?

Regards,

Sumit

[bhavesh_kantilal]

Use a BPM with Collect Pattern and the corrsponding Correlation and N:1 mapping.

regards

Bhavesh

[Former Member]

Hi ,

<i> Ztable there is one ITEM field and other fields based on the three files</i> . As he mentioned earlier..

I think BPM can be avoided.. Not sure if the new suggestion is the best way but worth a try..

Since you have the key field with you ie ITEM in all the 3 files . You can use MODIFY statement in the proxy or RFC code. So you actually dont need wait to collect data from the 3 different files .

Bhavesh what do you think of the above idea?

Regards,

Sumit

[bhavesh_kantilal]

Sumit,

The data is available in 3 source files . How would the R3 system get this data?

I think BPM is needed considering that the 3 source files would contain the data and the RFC / R3 system would need all this data to do the further processing.

Regards

Bhavesh

[Former Member]

Hey

since he has three different sender Files,he will be creating three abstract interfaces and will be using them in BPM.

if some how he can create a generic message interface which can handle any file,then i guess it might be possible without BPM,but i guess without that it might not be possible.

Thanx

Ahmad

Message was edited by:

Ahmad

[Former Member]

Hi bhavesh,

<i>The data is available in 3 source files . How would the R3 system get this data?</i>

Let us consider with an example..Suppose in the structure there are 6 fields to be updated. Now what the requirement is that theres a key field which is present in all the 3 files ie ITEM. Let us also consider that out of the final 6 fields which are to be updated in R/3 each source file contains 2 fields along with the Key Field ie ITEM.

We can create 3 different interfaces .. And which ever field is required for the interface update data in ztable in R/3 with those 2 fields. We already have the ITEM field ie Primary Key.

Lets assume we get the first source file ..we have the key field and 2 other fields which needs to be updated in R/3. the interface updates the data in R/3.

Next time the second source file comes..we would again have the key field as well as 2 other fields.. the interface updates the existing entry(with the help of key field).

similary for the 3rd source file the update happens.

Regards,

Sumit

[bhavesh_kantilal]

Sumit,

The fact is that ,

1. You are creating a ZTABLE on your R3. Beleive me in real time cases your R3 guys would never agree to it. They would not want customizations .

2. It makes more sense to correlated all these data in a single point - XI than at multiple points and programatically on youtr R3 system.

I agree that this is a solution that makes sense but at a bigger picture, the answer you will get from your functional / clients is why use a Middleware if you want to do all this Programatically on R3.

Regards

Bhavesh

[Former Member]

Bhavesh,

thought wud be gud if we can avoid BPM.

Agree with you.

Thanks,

Sumit