Limitations
Manual Merge
Rulebase does not work with child repositories and relationship attributes. When a child record is added from the source or other matched records, then the validation process is executed, the newly added child records get validated, but the existing record does not get validated.
Index Entities
The Matching across repositories is based on the indexing of multiple repositories in a single Netrics table as a composite entity.
This approach of using a denormalized table for indexing has several fundamental disadvantages, such as replicating data, requirement of more memory for too many record bundles, and so on. These disadvantages counteract with the fast searches on the denormalized table, which do not require any joins.
Therefore, you must carefully specify the composite entities and limit their complexity. The following are the various limitations that you must be aware of while specifying the index entities:
- Composite index entities cannot be defined across a self relationship.
- The repositories in a composite index entity must be specified in a linear way. The index entity does not allow two relationships in one repository, which is implemented in the Index configuration file. Each repository tag inside the index entity can only have one associated relationship.
- Relationship attributes cannot be used in matching expressions. However, the Text Search web service allows specifying relationship attributes for a search criteria.
- Defining an index across reverse relationships is not possible. Similarly, matching and text searching across reverse relationships are not supported.