Master data transfer - ECC to APO

[Former Member]

Hi,

I need to understand how transfer of master data from ECC to APO takes place?

There is a procedure to setup config for change pointers through T Codes BD61, BD50 & then generate & activate the relevant integration model.

There is another Transaction CFP1 which evaluate and send the change recordings. What is the difference between this and above given approach?

regards,

Mohit

[rajkj · Accepted solution]

Hi Mohit,

It depends on your business requirement i.e. whether you need the master data changes to be transferred from ECC to APO immediately or periodically. Use t.code CFC9 to configure the mode of change transfer of master data.

But, the prerequisite is to have an active integration model and activation of ALE change pointers (t.codes BD61 and BD50).

If you select periodic transfer in CFC9, you need to use t.code CFP1 to send the changes to APO. On the other hand, if your choice is immediate transfer, then CFP1 need not to be executed.

Thanks,

Rajesh

[Former Member · Accepted solution]

Hi Rajesh,

Thanks for responding.

Does it mean that if we select ALE(1) then apart from other settings we have to use CFP1 &

if we select BTE(2) then CFP1 is not required.

Please confirm on this.

regards,

Mohit

[rajkj · Accepted solution]

Hi Mohit,

That's correct. We use BTE for transferring the transactional data changes. There is no periodic transfer in this category and hence, CFP1 does not support this kind of data transfer.

Whereas ALE and master data change transfer, you have the flexibility to choose between periodic and immediate. CFP1 will be useful for periodic transfer.

Thanks,
Rajesh

[Private_Member_738444 · Accepted solution]

Hi Rajesh/Mohit,

I have been using CFM1, CFM2 & CFM3 along with occasional usage of RIMODINI for CIFing my master data as well as t-data from R3 to APO in my testing exercises in development systems.

However in production systems we have background job runs (daily) for CFM1 & CFM3 for daily updating of the integration models & thereafter we have these always active for instantaneous transfer of t-data between R3 to APO & vice versa.

However I fail to understand your info on BTE & ALE mode of CIFing. Pardon my ignorance on the same & please elaborate how are these different from CFM1/2/3 & RIMODINI?

Kumar

[rajkj · Accepted solution]

Hi Kumar,

CFM1 useful for generating an integration model.
CFM2 activates the model generated by CFM1 (and transfers all the data of IM to APO) and the corresponding program is RIMODINI.

CFM3 activates the model and transfers the delta to APO.

The active integration model is the prerequisite for data transfer to APO.

The CFM1/2/3 interfaces define run time objects (based on the master data collected at the time of execution) in ECC and APO to enable CIF transfer.

In this thread, we discussed about immediate and periodic data transfers to APO using ALE change pointers (master data) and BTE (transactional data).

We need to schedule background jobs (CFM1 and CFM3) to run periodically during off peak hours to ensure that any new master data created in ECC (still with no run time objects) need to be collected and CIF'ed to APO (that means run time objects will be updated with new master data).

I hope this helps you.

Thanks,
Rajesh

[Former Member · Accepted solution]

Hi Rajesh,

There are 2 scenarios:

1  Creation of new master data

2  Changes in the existing master data

The procedure you told in above is for change in master data.

For the new master data we have to follow the procedure of generation & activation of integration models.

So does this mean we have to set up background jobs for both scenarios separately?

regards,

Mohit

[rajkj · Accepted solution]

Hi Mohit,

The auto or periodic transfer of changes will work only for master data part of the active integration models. If there is a new material, the RTO needs to be updated. This is done through t.code CFM3 (deactivation and activation of the model in background).

If you wish to use periodic transfer of changes, then you need to have 2 sets of batch jobs:

1. Deactivation and activation of integration models

2. Periodic pull of changes

Or if you use change pointers, only deactivation and activation of integration models job would be sufficient.

Thanks,
Rajesh

[Former Member · Accepted solution]

Hi Rajesh,

Thanks for the reply.

I am sorry..I could not understand the term RTO..please expand the same.

Secondly as per my understanding activating change pointers means that it is meant for periodic transfer. Am i correct on this?

regards,

Mohit

[rajkj · Accepted solution]

Hi Mohit,

1. RTO - Run time objects or run time version. Assume this as a snapshot of master data collected when a model is generated and activated.

2. As mentioned earlier, change pointers activation is required for both periodic and immediate master data change transfer. However, you can choose the timing of data transfer i.e. immediate or periodic using t.code CFC9. With periodic transfer, the change pointers will be collected and sent to APO whereas immediate transfer will send the corresponding change pointer to APO as soon as it is recorded in the ECC system.

Thanks,

Rajesh

[Former Member · Accepted solution]

Hi Rajesh,

Sorry for a late query...

For online transfer of master & transaction data we have to generate & activate integration model on regular basis.

What is frequency for the same normally kept?

regards,

Mohit Goyal

[rajkj · Accepted solution]

Hi Mohit,

No issues.

Typically, in CPG industries, it's once per day where as in manufacturing industries (where changes are not frequent) it might be once per week.

Thanks,
Rajesh