PID Numbers Crossed After a Merge
I don't know if this the right place to submit this question, but I can't find anywhere else to post it, so here goes:
I was doing some merges and noticed that PID# 9J3Z-3HG (Margaret Ethel Dillon) has an unusual sealing-to-spouse message on her Ordinance page. The message reads:
“These people do not have a couple relationship but are the parents of a child. If they were married or lived together as husband and wife, add a couple relationship to link them as spouses, so that you can request the sealing. If they were never married or never lived together as husband and wife, do not have them sealed.” This message applies to a PID# LT9X-97N.
When I open another Family Search window and do a “Find” for LT9X-97N, the system brings up Margaret’s husband, William Grant Pontis (PID# LDY8-Y59). Yes, when I do a find for LT9X-97N, the system brings up a different PID# instead – LDY8-Y59.
After much searching, I found that one of Margaret Dillon’s children (PID# LT96-5Q6), Marjorie Louise Pontis, is probably part of the problem. You will see in the notes that I merged duplicates for Marjorie Louise Pontis: LT96-5Q6 with PID# L744-LL7. When I attempt a Find for L744-LL7, it brings up LT96-5Q6 instead.
Marjorie had been separated from the other children with a duplicate set of entries for the parents, Margaret Ethel Dillon and William Grant Pontis, so I merged the parents and brought Marjorie into the family. I’m not sure what happened that caused the sealing-to-spouse message on the Ordinance page for Margaret (9J3Z-3HG). Nor do I understand why a search for one PID number brings up a different one. Something got crossed. Can you figure this out? That sealing message will drive people crazy trying to figure it out. Thanks.
Paul W ✭✭✭✭✭
Maybe the "Temple" category would have been more appropriate - as ordinance work is not a subject a lot of Community participants are familiar with. No matter, you will still probably receive an appropriate response from a Church member who has knowledge of such matters.2
Gordon Collett ✭✭✭✭✭
Since this is more about merging and ID numbers than anything else and it does look like there might be a new bug somewhere, this question probably is most appropriate here.
First off let me review what is going right.
A recent update in the the FIND routine has made it so the current existing person represented by an ID is always brought up in the result which I think is very reasonable. The behavior you are seeing with the ID numbers is working correctly in accord with this model. As you describe above, when you use FIND to search by ID, if that ID has been deleted in a merge, you get the survivor in the merge.
For example, if AAAA-AAA has been merged into BBBB-BBB who was then merged into CCCC-CCC who was then merged into DDDD-DDD so that only DDDD-DDD now exists in Family Tree and represents all four of those records, using FIND to search for any of those four records, brings me to his current record, DDDD-DDD. This is actually probably what most people really want unless they are specifically working on correcting an incorrect merge.
We can still access those deleted record by going through DDDD-DDD'S Change Log or by using the search box at the top of the Recent's menu.
Now to what is not working right.
Either a new bug has crept in or the merge for William Pontis did not go right. Margaret Dillon only shows one husband on her detail page but she shows two husbands on her ordinance page:
These two records were merged. But apparently only half merged because only the survivor should be showing here.
Whether this is a bug that has crept in or lightning storm/power surge/server glitch data corruption is hard to say.
What I would be inclined to do is to restore LT9X-97N, restoring William G Pontis, then merging the two records again.
Remember, you can only get to LT9X-97N now by using the Change Log and clicking on the name of the deleted person:
Or using the Recents menu and pasting his ID number in that search box:
If you get the same result after restoring then merging again, then this is a bug a moderator needs to get passed on to the engineers.
On to another topic that I noticed while looking at your problem.
I see in William's Change Log that you have dismissed a number of hints for him. These are most easily seen in the Research Helps - Dismissed section:
FamilySearch indexed sources are all individual-centric in that there is a unique source created for each person in a source record, one each for each person mentioned in the record.
You see see these individual source when a source is evaluated in the Source Linker. To take the first of the above four sources as an example:
Each line on the left is a different, unique source for only the person listed there and no one else. This single marriage record was used to create six sources, one for the groom, one for the bride, and four for their parents. Five of these sources are correctly attached to the correct, corresponding individual.
You have declared the sixth source as "not a match" which is incorrect. The biggest problem with doing so is that by dismissing this hint, stating "not a match", and not attaching this source, you have just told the Hints routine that William Grant Pontis is not the father of Paul Pontis and is not the husband of Margaret Ethel Pontis. The Hints routine looks at all the family relationships for a person, all the existing sources, and all the rejected sources on a person to find more sources for you. It will now very helpfully do as you told it to and disregard other sources for William that show any relationship to Margaret or Paul. The more correct hints/sources you mark as incorrect, the fewer new correct hints you are going to get on these people.
You really should go back and attach those for sources which are indeed for William to William's record so you don't confuse the Hint engine.4
Hi Gordon -- Thanks for your very speedy reply and your clear explanations. You're really good! So, FIRST, I did unmerge PID#s LDY8-Y59 and LT9X-97N, as you suggested. After unmerging them, I carefully reviewed the notes looking for anything that might explain the problem with PID# 9J3Z-3HG (Margaret Ethel Dillon -- the wife) two husbands showing on the ordinance page. I couldn't find anything that seem inappropriate, so then I remerged LDY8-Y59 and LT9X-97N. After doing the merge again, I checked the wife's page, and there's still a duplicate husband with PID LT9X-97N. One thing I don't understand is why LT9X-97N on the ordinance tab for Margaret Ethel Dillon (PID# 9J3Z-3HG) shows as "[Unknown Name]." I checked the notes for Margaret, for her husband, William Grant Pontius (PID# LDY8-Y59), and for their daughter, Marjorie, who was the only child originally attached to LT9X-97N. I couldn't find anything that would account for "[Unknown Name]" UNLESS it occurred when the account for LT9X-97N was originally set up. The very first entry that was made in his notes shows "Person Added" with no name showing. Just a guess. So, I don't want to keep performing attempts to fix this because I really don't have any great ideas, and I would just be adding a bunch of confusion to the paper trail. If you have some data guys who could take a look at this, that would be great. The ultimate goal is to remove the extra spouse "[Unknown Name]" from Margaret's "Ordinance" tab.
SECOND: I'm glad you raised the issue about processing hints as "Not a Match." This issue is a pet peeve of mine. I have many families on my English lines who had 10 ... 12 ... even 14 children. When you start adding all the records for the children to the parents' source pages, you can end up with 100 or more sources that do NOT actually uniquely identify the parents. For example, a child's birth certificate often has the parents name on it but often has no additional information that identifies the parents. If you have a father named "Tom Smith" or a mother named "Mary Jones," you have no idea WHICH Tom Smith or Mary Jones in the history of the world that child belongs to. That child's birth certificate often does not uniquely identify a parent, other than providing a name. The name DOES narrow down the search, but it's not a unique identifier. A name alone doesn't tell you who the parent is and the name can easily be seen on the "Detail" page for the child, so why would it even be necessary to put in the sources for the child. The parent isn't uniquely identified until you find birth marriage and/or death specifics. Adding all the records for children to the parent's pages just because there is a name match to the record clutters up the source page and makes it infinitely more difficult to examine the records that actually DO apply to the parents. I've never had a problem with this. The system seems to continue to scan for records because I sometimes have the same record that I called "Not a Match" come back as a hint again. I continue to receive hints for all family members. It's really cumbersome to go onto someone's "Source" tab and find 40 or 50 sources, haphazardly entered with no attempt at organization, and not necessarily applicable to the actual person the source page belongs to. Further, the titles of the sources often do not provide specific information as to who or what the record is. Some sources simply say they are "Parish" records. Is it a baptism? Marriage? Burial? So you have to wade through all these extraneous records trying to find the information you're looking for. So ... that's why I rejected some of the hints. In my opinion, it makes the "Source" pages more focused.0
Gordon Collett ✭✭✭✭✭
1) I'm just another user so can't do anything about checking the data structure as to why the merge is not working the way it should. You'll have to wait for a moderator to see this post and send the problem off to the right team, one that can tinker under the hood of the database.
The reason LT9X-97N, the former William G Pontis, shows as "Unknown" on the ordinance page is because he has been deleted through the merge. I assume that when you restored him the ordinance page showed his name again? If if didn't then something is really messed up in the underlying database structure. The problem is the couple relationships between Margaret and William G and Margaret and William Grant are not merging the way they are suppose to.
I fully support just waiting to let someone in data administration fix this so Williams's Change Log is not a complete mess. There are limits to what you and I as regular users can do here.
(But I will admit I like to do as much as possible here on my own and find it fascinating to see just how much we can do. So if this was my problem - and it is not - the two other things I would try would be to restore William G and try merging in the reverse order, leaving LT9X-97N as the surviving record. If that didn't work, then I would again restore so that I have the two Williams, then remove William G LT9X-97N from being Margaret's husband and Marjorie's father, make sure Marjorie's only child-parent relationship is with William Grant and Margaret, make sure everyone's ordinances pages look fine and there is no sign of any residual relationships that should not be there, then merge William G as an isolated person without any family relationships into William Grant. The I would be very disappointed when none of that worked. So just wait for a moderator to chime in here. It may not be until Monday since I hope they get days off.)
2) For the past ten years people have been complaining on the various FamilySearch feedback boards of this interesting problem of too many sources. A fascinating contrast to the main complaint about other online platforms regarding so many person records not having a single source. Requests for some way to organize them, group them, or hide redundant copies are made on a regular basis. So you are not alone in your complaint.
Regarding the haphazard appearance of the source list at times, have you found that using the option to have the sources sorted chronologically helpful?
Personally I generally like to just drag the sources into a Custom Order that I like unless someone else who is watching the record has clearly already dragged them into an order that user likes. Regarding the vague titles, I agree it would be nice to see more specific information but I suspect the automated routines that generate the title can't handle that. All the titles are is the name of the person, whether the record is for that person or that the person is mentioned in someone else's record, and the title of the database it came from. I use the date listed to tell me what the record probably is. And, of course, we can edit those titles if necessary to provide fuller information if we want to take the time to do that.1