Sometimes despite the data not having matching attributes, you know two records belong to the same person or organization. Senzing provides the ability to force such records together simply by utilizing the TRUSTED_ID feature. TRUSTED_ID will force together records that share the same type and value regardless of the data. The records can be from the same or different sources and data loads.
Different TRUSTED_ID features will prevent records from coming together automatically but that behavior can be changed.
Consider the following two records:
1 ,SKIPPY ,1960-01-01 ,5551212
2 ,BOB JONES,1960-01-01 ,5551212
Senzing would detect they share the same date of birth and phone number but could not tell you that Skippy and Bob Jones are the same person at this time and with this data. Maybe these are twins or a misassigned date of birth.
First, if you have data that informs you these are the same, add that data to the system. It is always better to give Senzing more data to support it automatically making decisions than overriding it's behavior.
You can use the TRUSTED_ID feature which has TRUSTED_ID_TYPE and TRUSTED_ID_NUMBER attributes. In this example we use 'FORCED_SAME' as the type and utilize the RECORD_IDs to create a unique entry for the value. For good measure, we will prepend SAME to the attributes as the usage type to separate it from other similar values.
1 ,SKIPPY ,1960-01-01 ,5551212 ,FORCED_SAME ,R1-R2
2 ,BOB JONES,1960-01-01 ,5551212 ,FORCED_SAME ,R1-R2
Reload those records and you will see your decisions will have been applied to the results.
- You can add the new attributes to the records by calling getRecord() in the API, add the new key/values to the record, and then call replaceRecord() to reload. Senzing will do the rest. You could do the same thing to clear the FORCED_SAME override.
- More than two records can share the same TYPE/VALUE pair. In that case all the records that share the value will come together. This can be helpful if there are a group of records you want to make sure stay together despite the data.