cancel
Showing results for 
Search instead for 
Did you mean: 

Mass Import, unable to use Suppress Unchanged Records

Former Member
0 Kudos

Hi everyone,

I am looking to use the suppress unchange records option for syndication. Currently when we do a large import and have a matching remote key we use the UPDATE ALL MAPPED FIELDS for import (even if the record is unchanged). However this option does not work for the Supressing Unchanged Records, probably because we UPDATE all records that have a matching remote key. It seems that the Syndicator thinks that these updated records have changed, when the record has not actually changed. Is there any solution to only syndicate records that have actually changed and have not simply been updated by the import?

What are your thoughts?

Thanks,

Matt Winchester

Accepted Solutions (1)

Accepted Solutions (1)

Former Member
0 Kudos

Hi Matt,

this is the normal behavior of syndication, when you import a master data record using Import Manager, this means an update for Syndicator. Therefore, the "Supressing Unchanged Records" option does not suppress that record!

Let me try to explain you a possible workaround. You need to compare the old value with the new one for each fields inside the record, and if one of the new values is changed then go ahead with syndication. You can do that using the check-out functionality, therefore put all the master data records in check-out status after import process just using an MDM workflow containing an expression that checks if there is a difference between the new value and the old one. If there is a difference the workflow will syndicate that master data record using the syndication step, otherwise it does not syndicate it. Closing the WF the master data record cames back in check-in mode.

The expression should be something like this:

IF COMPANY NAME = [ORIGINAL].COMPANY NAME ..................

The above compares the new value with the old one.

Hope this can help you.

Best regards.

Vito

Former Member
0 Kudos

Thanks for your quick response. My concern with this solution would be performance related. If we have 60,000 records individually running through a workflow each day, we will most likely have performance issues. Any thoughts?

Former Member
0 Kudos

Hi Matt,

you post is very interesting. I don't think that you should encounter performance problems, but you just need to take care to delete from MDM the completed WF, this is a standard operation that should be done consistently in order to lighten the MDM and to improve the performance, do to that you can use a simple scheduler task (MDM Java API are better because actually are more powerful then MDM Abap API).

But, let me say, you can also change the "update all mapped fields" in import manager, but I don't known your business requirements, probability you need that configuration. Furthermore, if it is fundamental to do not syndicate the unchanged master data you could also think to implement some logics on a middleware like PI, i.e. before to load data in MDM through Import Manager, PI should call MDM in order to undesrtand if an equal master data already exist, in that case PI could decides to do not create the XML file for that master data, therefore to do not update the unchanged master data... then the work is done: the unchanged master data will not be syndicated.

It is clear that you could introduce a logic at PI layer between MDM and the destination back-end, in this case you syndicate all master data records but PI will be able to filter the unchanged master data records.

An important consideration: using the previously solutions you don't have the components decoupling anymore....

Hope this help you.

Best Regards.

Vito Palasciano

Edited by: Vito Palasciano on Nov 21, 2009 7:15 AM

Answers (0)