Rank
After you configure all the declarative rules, rank them in the order in which you want to run them for matching record pairs. The declarative match rules are run on one record pair at a time.
First, the declarative rule with rank 1 is used to process a record pair. If the record pair meets the conditions set by the match rule, the merge strategy for the rule is applied to the record pair. The processing for the record pair is complete.
If the record pair does not meet the conditions set by the match rule, then the rule that is next in rank is used to process the record pair.
If the record pair does not meet the conditions set by any of the ranked rules, the unranked rules, which are the rules with the threshold-based merge strategy, are applied to the record pair. All the unranked rules are applied in succession, and the outcome is assessed to make a unanimous decision.
The following table lists the actions on record pairs for the unranked, threshold-based rule outcomes:
Unranked Rules Outcome | Action on Record Pair |
|---|
Majority of the rules have the same outcome | Majority outcome is applied. For example, if the majority outcome is automated merge, then the record pair is queued for automated merge. |
Tie between manual merge and automated merge | Queued for manual merge. |
Tie between automated merge and skipped merge | Queued for manual merge. |
Tie between manual merge and skipped merge | Skipped and will not be matched again. |