CiviCRM Community Forums (archive)

hi,
i think that the reason no dupes were found is because the fuzzy dedupe rule you are using is examining the entire organization name instead of just part of it. if you edit that rule and set the "length" for organization name to e.g. 8, you will get a match for the two records that begin with "The Cardinal". you will not, however, get a match for the "cardinal wiseman" record because there is no way (that i can see) that it will ever match the two records that begin with "the".

hope this was helpful,
cynthia

alanski,
i do hear your frustration but in this particular instance i am not sure i agree.

under the best of circumstances, deduping data is a daunting task. that is why some organizations pay thousands of dollars for software that is devoted to that task. and even with that special software nothing, in my opinion, replaces the human aspect of eyeballing data and handling each instance with care so as not to merge records that are not really duplcates. i think this is true regardless of whether or not the software is "open source".

the specific examples you listed might be able to be handled by the code if it were modified say, to ignore the word "The". but there are hundreds of other examples where it would be difficult to apply logic to determine if records were duplicates. this, in my opinion, is why standardization of data entry is so important. having said that, i realize that standardizing data is easier/doable if data entry is done by staff and not done directly by "members".

i hope you can figure out a way around these issues. post your ideas and if the community can help you i am sure it will.

cynthia

DOnt get me wrong Civi is awesome . But there are just some very painful issues with it {eg CHange a button text comes to mind....eg write a Joomla plugin to hijack the buildform event and manipulate a button name value - cmon!}

I havent fully explained case scenario behind this 'whinge'

Membership sign ups on behalf of create a new record 'Cardinal Wiseman' which is a duplicate of  existing record 'The Cardinal Wiseman School'

The original record has loads of other data attached to it, geo stuff and custom data.

So the good admin wants to dedupe memberships/org records.
What are the choices?
Edit membership/trasnfer to another Org contact? Nope cant do that

Only way I can see without development is to dedupe and merge.

But the dedupe can't match the two words that are the same for three contacts.

Ergo duplicates exist and cant be changed without heavier code solution.

My 'solution' to this is to ajax autocomplete the Org onBehalfOf field which should catch most but not all before dupes are created.

But it doesn't guarantee resolve the issue.