Data Cleansing 

Use

The data cleansing process step (MDM_CLEANS) processes (inbound) master data records that have just been loaded to the master data server (MDS) and master data records that already exist on the MDS (downstream data cleansing).

The data cleansing process step contains the following substeps:

·        Matching of master data objects

·        Merging of master data objects

Prerequisites

You have performed the following IMG activities in Customizing for Master Data Exchange (MDM-MX):

·        For matching, IMG activity Define Settings for Object Types of Master Data Server (for each object in the Matching area).

·        For merging, IMG activity Define Object Types of Master Data Server (for each object in the Staging area).

Object-specific classes determine what is carried out in this process step in detail. SAP delivers classes. You can develop customer-specific classes based on these classes and assign corresponding object types.

For the "business partner" master data object, you can use function module MDM_BUPA_MERGE_EXECUTE as a copy template for implementing customer-specific merging logic.

For more information, see the function module documentation.

Features

Matching

Matching uses SAP Content Integrator (CI) to check whether any loaded master data records are semantically identical. This check is rule based and runs automatically. As a result, CI returns an ID mapping to the MDS, which provides information about which data records are identified by the system as identical.

Matching for an object from an MDC only runs the first time you load the object.

If the rules cannot identify a definite match, you must manually make a decision. Data cleansing cases are created for these objects on the MDS. However, you can define company-specific rules to control the automatic matching decision.

For more information, see the IMG for the data cleansing Business Add-In BAdI: Control Results of Matching Process.

The data cleansing cases created by the system are processed by master data experts with the user interface for Data cleansing cases.

The data cleansing cases are saved as inactive, and are only visible for the user once matching has been completed and the automatically recognized duplicates have been merged.

 

Merging

The merging of master data records uses the staging database. This is done on the basis of rules, automatically in the background, or manually. If the rules cannot decide on a merge, you must manually make a decision. Data cleansing cases are created on the MDS and can be processed with the user interface for data cleansing cases.

Possible process flows for merging are:

·        During matching, the system finds two newly loaded data records that belong together and are not known to the MDS. These two data records are deleted from staging in favor of a new object on the MDS and are not posted to the MDS (two old data records are deleted and a new data record is created). The records are retained in the MDCs.

·        During matching, the system determines that a newly loaded data record from an MDC and a data record in the MDS are identical. In this case, the data record in the MDS is updated with the data from the MDC.

·        In XI mapping, the system determines the corresponding MDS object. The system merges the loaded data record with the MDS object. There is no subsequent matching on the MDS.

Once merging is completed for all the objects in the process chain, or you have processed and released all data cleansing cases, the status of the process step is set to finished and the process chain continues. Changes to the objects are not active until the post process step has written the changed or newly created objects from staging to the database of the master data server. The objects that were recognized as duplicates are given an archiving indicator and a distribution lock is set for these objects.

You can also define company-specific follow-up actions for duplicates.

For more information, see the IMG for the data cleansing Business Add-In BAdI: Control Standard Behavior of Data Cleansing Cases.

For more information about the effects of merging at object level, see Merge Master Data Objects.

Configuration

Create a variant of this process step with the following configuration parameters for each process chain:

·        Object type

Object type that is sent for matching (for example, BUS1006)

·        Maximum number of objects

Defines the maximum number of objects in a process in the matching tool (for example, SAP Content Integrator (CI))

·        Maximum number of processes

Defines the maximum number of parallel processes in the matching tool (for example, SAP Content Integrator (CI))

·        Trace level

Defines the type of information that is written to the log file.

The data cleansing process step has its own monitor.