Steps in Key mapping

Question

Hi all,
i have gone through many forum's regarding key mapping.
i can see that the following are used..
--> remote systems
--> key mapping option in table-field properties
--> key mapping option in table properties.
-->key generation(none, range, qualified range) field in remote system properties.
But never understood like in which order we need to define all the above, what is the relation b/w them.
i have gone through some theory in sdn regarding key mapping but unable to get the logic..
can any one one show me with 2 records in 2 tables so, that i can clear my confusion.
Thanks in advance.
Regards
phani

Answer 1

Hi Phani,
first you should understand the basic idea of key mapping. MDM usually wants to consolidate master (objects like vendors, customers, and so on) and/or reference (lookup tables like countries, regions, and so on) data. The idea is to load this data from various remote systems into MDM, compare and match the data, and in some cases merge multiple records into a single one. Let's take an example.
Image you have attache two remote systems to MDM. The first is an ERP system, the second a CRM system. Both system contains countries as reference data. In ERP, your country India looks like
Name = India
Code = INDIA
whereas in CRM it looks like
Name = India
Code = IND
It is obvious that the country is the same, but the systems use a different internal code. Now you load this data into MDM. What do you do? Do you want to create two records having different codes? Or do you want to "merge" the indentical records into a single one. The common decision in MDM is the second one.
But what happens now if you send customers from ERP and CRM to MDM? Usually the Customer message contains the Country CODE but not the Country NAME. From ERP the message could look like:
Customer = Mr. X
Country = INDIA
whereas from CRM the message could look like
Customer = Mr. X
Country = IND
Surely you want to map the different code in both case to the same country "India". That's why you use key mapping in MDM. Key Mapping allows MDM to store different codes for the same data. So in MDM you store just one record, but if you receive and/or send data from and/or to remote systems, you can always ensure that you store and/or send the correct codes.
Hope that clarifies your doubts about key mapping in general?
OK, now regarding your question about "Update NULL fields only":
In Import Manager you have several options for data import. If you import for the very first time (for RS1), your table in MDM is usually empty. That why you need to "Create" the records. If you import the second time (for RS2), your table in MDM already contains data. Depending on the matching field for import, several options are now possible. If Import Manager is not able to match the record you want to import against any other existing in the repository, you have to "Create" the new record. If Import Manager is able to match the record that you want to import with a record of your repository table, you can choose what happens next. "Update NULL fields only" compares the record you want to import with the existing values in MDM. In your case, following info is already part in MDM:
Country = India
Code = INDIA
Remote Key RS1 = INDIA
Remote Key RS2 = NULL
Now the file you want to import contains the following values:
Remote System = RS2
Country = India
Code = IND
Import Manager identifies that only "Remote Key RS2" is NULL. All other fields already have values. As you choose to "Update NULL Fields only", only Remote Key RS2 gets a new value. If you choose a different option, e.g. "Update All Mapped Fields" Import Manager would update the Country and Code fields, too.
Best regards
Michael