In this scenario, you need to create and define a match rule
MatchCustomer to match the staging data records that belong
to the Customer entity based on the fname
and lname fields.
In MDM, match rules are used to decide whether two or more data records match, and how to
handle them if they do.
-
In the MDM Repository tree view, right-click
Match Rule and then select New
from the contextual menu.
-
In the dialog box that opens, define a name for the new match rule.
If needed, enter information in the Purpose and
Description fields to better describe your match
rule.
-
Click Finish to close the dialog box.
The newly created match rule is displayed under the Match
Rule node. You need to further define the characteristics of the
match rule in the Match Rule Editor that opens.
-
In the Record linkage algorithm section, select
T-Swoosh.
You can use the T-Swoosh algorithm to find duplicates and to define how two
similar records are merged to create a master record, using a survivorship
function.
-
In the Match and Survivor section, define the criteria to
use when matching staging data records.
In this example, add two match keys Firstname and
Lastname, select Jaro-Winkler
as the matching function, set both thresholds to 0.8, and
select Longest (for strings) as the survivorship
function.
-
In the Default Survivorship Rules section, define how to
survive matches for certain data types: Boolean, Number and Date.
If you do not specify the behavior for any or all data types, the default
behavior is applied.
Once you define the match rule, you must attach it to a specific entity of
a data model.
You cannot deploy a match rule directly to the MDM server. Rather, match
rules are deployed along with the data model to which they are
attached.
In this example, add two match keys Firstname and Lastname, select Jaro-Winkler as the matching function, set both thresholds to 0.8, and select Longest (for strings) as the survivorship function.