Importer should skip duplicate attribute values set in splitters
When metadata is mapped via the attribute parameter within, for example, an entitySplitter the importer adds the attribute metadata again with each relationship. This is good practice in general because each new relationships may be linked with new attribute data. However, a lot of times the attribute data is exactly the same as what's already there, ie. the entity’s phone number is already cataloged and we don’t need it again. What do you think about checking to see if the exact value is already present in the target attribute during the merge and if so, skipping the addition of the duplicate data?