The Duplicate Constituent Report takes days to run and then returns thousands upon thousands of records that are not true duplicates preventing de-duping of the database with this tool. I would advise anyone who has not upgraded to 7.92 to de-dupe their database before installing the upgrade or until this issue is resolved.
This patch 6 enhancement:
"Missing address information will reduce the likelihood of being considered as a potential duplicate"
doesn't seem to be working for us as the duplicate matching (during imports) are still throwing up matches to people who are on our database with completely blank addresses.
Is there any way to perhaps base the duplicate checking on a user-controlled subset of the database (we could define this by setting a standard query in business rules).
At the moment the duplicate matching in imports is producing far too many false positives [really obvious ones, as well] and is by far the most time-consuming step during imports.