CiviCRM Community Forums (archive)

Having just imported a little under 1000 new contacts and placing them all into a group specifically for the purpose, I ran the "Find and Merge duplicate contacts" as I new there were several contacts in the import data that were the same person, but with different email addresses. I set up a dedupe rule to find matched based solely on the name fields, and selected my newly created group to limit the search only to the newly imported contacts.

I was surprised to see that the search results included contacts not in my group of recently imported records. It looks like CiviCRM is searching the entire database for matches based on my dedupe rule.

Is this expected behaviour? The documentation (http://book.civicrm.org/user/current/common-workflows/deduping-and-merging/) states:

You can then select to search all contacts for duplicates or to search only a particular group.

To my reading this means that the search for matches should be limited only to those contacts within the specified group.
So, is the code doing the right thing and the documentation incorrect or misleading. Or is the documentation correct and CiviCRM doing the wrong thing? If Civi is searching all contacts, which looks to be the case, then there is no point in it offering to limit the search to a specific group. Consequently I would say this is a bug.
This is using version 4.4.3 on Drupal 6.

Best wishes
Graham

Hi Coleman

It looks as if at least one of the pair is within the group. Running the same rule without limiting to the group gives a larger result set, so limiting the search to the group is clearly having some impact on the result, just not the one I was expecting, based on the documentation and my assumption as to how this limit would work.

I can see the value of an approach that would look only for matches where at least one of the records is within the specified group. Assuming this is the behaviour that is being seen, perhaps what is needed is a further limiting option whereby we can limit searches even more by ensuring that both records need to be within the specified group?

In any event, the documentation needs to be clearer IMHO.