Data Merge - Master and Reference Data

[Former Member]

Hello All

We have a question regarding the merge process of out Reference data and also our master data.

We know that we have duplicate reference data values across our multiple backend systems. Do we need to merge these values prior to loading the main table records? If we do not merge prior to loading master rereads will MDM prevent you from merged the reference data value associated with these master data records?

With out duplicate reference data records we noticed also that some of the duplicate records are identical.

e.g. System 1 has Material Group value 1000 and

System 2 has Material Group value 1000 also

When we load master data records for each of the systems how can we distinguish which is the correct material group lookup value ( if we do not have consolidated reference data.)

Has anyone else had the same issues?

All help is appreciated

[michael_theis]

Hi ccr,

I'd suggest to do a reference data consolidation before importing master data. I know that there's no possibility to match and merge on look up tables, this is really a pity. Nevertheless it makes life easier if you have consolidated reference data.

Each look up table containting reference data should be key mapping enabled. The reason for this is the example with the material groups you mentioned. With key mapping enabled, you can store the material groups system dependently. Imagine you have Mat Group A in System A with key 1000. You've got Mat Group A (same name) in System B with key 2000. You have a duplicate Material from System A and B both having Mat Group A. f you syndicate this back to the source systems, Syndicator is able to pick the correct Mat Group Key out of the key mapping.

This scenario explains, why it is so important to consolidate your reference data. Depending on the match criteria during the import of the reference data, you might have one single record having two different keys (this is the result if you match via Mat Group Name) or you might have two duplicate records having only one key each (this is the result if you match via the Remote Key).

Kind regards

Michael

[Former Member]

Hi Michael, I am also in same situation... By saying

"do a reference data consolidation before importing master data"--- Do I have to do consolidation in Import manager? I mean by map field/values for check tables? If I am wrong than please let me/us know How?

[michael_theis]

Hi R.n

if you have multiple remote systems, you can do a first consolidation during the import. You can use the display fields, other text or code fields that may possibly match during the import. This can sort out excact duplicates as the Import Manager can only perform an exact match. In addition you should have a look at the look up tables afterwards in the Data Manager. There you can do a manually match and merge of the Reference Data.

Kind regards

Michael

[Former Member]

Michael,

We want to do this but the question I have is that do we need to complete all consolidation of reference data prior to loading Main table records...or can we do t both master and reference data on a continuous basis as part of an on-going QA program

We have some resource constraints in that we do not have business people available to make a decision of merging of reference data...but rather a MDM team will be formed in the future to assess data quality issues go forward

[michael_theis]

Hi ccr,

I think you could do it system for system. You should have imported all reference data of a remote system before importing any master data of the same system. Otherwise Import Manager will fail to value map the incoming keys to their correspoding look ups. Nevertheless, if you've finished loading reference data for one system, you could immediately start loading master data. In parallel you could now load reference data of a second system as merging this incoming data with already existing data won't change the remote keys of the exisiting data.

Kind regards

Michael