Open LIL717 opened 9 years ago
Hmm...do we know data in the records that are duplicated? It probably indicates a bug in the algorithm... On Apr 13, 2015 7:05 PM, "LIL717" notifications@github.com wrote:
On my husband's phone, the above message is logged dozens of times. Is there anything we can do with all those dups?
— Reply to this email directly or view it on GitHub https://github.com/ceaseless-prayer/CeaselessIOS/issues/92.
I can debug with some breakpoints to see more specifically what is happening.
Ok, here is an example of when this is happening. My husband has 5 contacts with the name Bobby Cox and 2 with Robert E Cox. They are all the same person and any one of them has one or two of the 3 possible emails for Bobby Cox. I doubt he is going to take the time to clean up his address book since he has 1800 contacts. No great answer here, but maybe we could provide an interaction in the People view where the user could tap some sort of "identify duplicate" button and when it is in that mode all the rows that are selected would be grouped into one. Of course there would need to be a way to undo that too... because we aren't exposing linked people in Ceaseless, but an accidental tap of Karen Cox in between Bobby Cox and Robert E Cox would link them.
Hmm...so if I were to elaborate the example:
Bobby Cox 1 email1 Bobby Cox 2 email1 Bobby Cox 3 email2 Bobby Cox 4 email1 Bobby Cox 5 email3 Robert E Cox 1 email2 Robert E Cox 2 email3
So we would expect to be able to automatically combine: BC1, BC2, BC4
And we could potentially combine BC3, RC1 BC5, RC2
And in this case we would end up with 3 Ceaseless contacts. Are we currently ending up with 7 separate Ceaseless contacts?
I think that is the case. I can verify that tonight.
Lori Hill Sent from my iPhone
On Apr 15, 2015, at 1:44 PM, Chris Lim notifications@github.com wrote:
Hmm...so if I were to elaborate the example:
Bobby Cox 1 email1 Bobby Cox 2 email1 Bobby Cox 3 email2 Bobby Cox 4 email1 Bobby Cox 5 email3 Robert E Cox 1 email2 Robert E Cox 2 email3 So we would expect to be able to automatically combine: BC1, BC2, BC4
And we could potentially combine BC3, RC1 BC5, RC2
And in this case we would end up with 3 Ceaseless contacts. Are we currently ending up with 7 separate Ceaseless contacts?
— Reply to this email directly or view it on GitHub.
Verified that the following happens: AB Bobby Cox w/ email1 = new Ceaseless Bobby Cox w/email1 AB Bobby Cox w/ email2 = new Ceaseless Bobby Cox w/ email2 AB Bobby Cox w/ email1 and email2 = new Ceasesless Bobby Cox w/email 1 and email 2 - this causes the error because they algorithm found 2 people that match, so it just creates a new entry because it doesn't know which of the first two to link to AB Bobby Cox w/ email 1 = another new Ceaseless Bobby Cox because now there are 2 entries with email1 and it doesn't know which one to choose
Randomly also found, but not really related to this issue: First: Robert Middle: E Last: Cox w/ email3 = new Ceaseless Robert E Cox w/ email3 First: Robert E Last: Cox w/ email3 = new Ceaseless Robert Cox w/email3
I think this issue can be addressed in an update, rather than in the initial version.
Okay I will move it to a new milestone.
On my husband's phone, the above message is logged dozens of times. Is there anything we can do with all those dups?