Change pointers and RFC

[Former Member]

Hi All,

Is it possible to implement change pointers without creating idocs, but with only RFC's.

Actaully i am goign through aproject document which says the interfaces are implemented using change pointers and BADI's.I was in a understandign that there would be idocs created to communicate the master data changes to other systems.

But one of the person who involved in the interface development said that there are no idocs related involved related to the interfaces, its all RFC's enabled.

Can anyone please let me know if change pointers can also be implemented using RFC.

Thanks in advcane.

[Former Member]

Hi All,

Is it possible to implement change pointers without creating idocs, but with only RFC's.

Actaully i am goign through aproject document which says the interfaces are implemented using change pointers and BADI's.I was in a understandign that there would be idocs created to communicate the master data changes to other systems.

But one of the person who involved in the interface development said that there are no idocs related involved related to the interfaces, its all RFC's enabled.

Can anyone please let me know if change pointers can also be implemented using RFC.

Thanks in advcane.

[Sandra_Rossi]

Hi,

I think the same, i.e. standard only generates IDocs (by scheduling RBDMIDOC program). So they probably have implemented custom code somewhere, possibly in BAdI IDOC_CREATION_CHECK which receives the IDoc data to send via importing parameter and tells back the system to generate the IDoc or not. So, here they may have code to process it separately, through RFC as they said.

Or maybe they send IDocs through RFC, and they forgot that there were IDocs anyway (or maybe they purge them immediately)

Sandra

[Former Member]

Hello,

FMs like BAPI_MATERIAL_SAVEDATA allow to store changed or new data in a remote system. So change pointers might be written for a material. A subsequent programme might read pointers and changed data and call one of those FMs.

However, idocs are safer for many reasons..

kind regards,

hp

[Former Member]

Hello,

yes this is possible.

In TA BD60 you maintain the function module that should evaluate the change pointers (per message type). Usually this function modules then create IDocs. But you can do whatever you want instead.

Best regards,

Oliver

[Former Member]

Thanks guys for your replies.

Can anyone give me the link/documet to change pointers, which would help me in understanding the concept of change pointers.

[Former Member]

Hi,

this is described sufficiently in SAPHELP - in a nutshell:

Changepointers - as the name says - point at changes (change documents). Therefore the creation is together with change documents (at the end of FM CHANGEDOCUMENT_CLOSE.

Chane pointers are created for ALE message types (tcode WE81). This does not mean, that any ALE message has to follow - there might be another subsequent action.

Writing of CP must be enabled with BD50. With BD52 the tables/fields are customized. Precondition is, that the data element of the table field is enabled for change documents and that change documents are created. BADI

BDCP_BEFORE_WRITE allows further refinement (or any other action triggerd by the change). Note that both - change documents and CP - are created in update task - so you have to activate update debugging to see it in debug.

Storage of CP is in table BDCP2 (since release 7.00 former releases use BDCP and BDCPS). Read pointers with FM CHANGE_POINTERS_READ and set them to processed with CHANGE_POINTERS_STATUS_WRITE. If more than aprox. 1 Mio. pointers are read an overflow may occur - use CHANGE_POINTERS_READ_MODE_SET to set an appropriate block size and CHANGE_POINTERS_READ_INFO_GET to stirr the further processing. All FM are documented sufficiently.

Have fun!

hp

Edited by: Holger Pakirnus on Sep 14, 2011 3:58 PM