Index Editor corruption of censuses: Indexing still lost
When trying to group households and make other edits to census records, the Index Editor used to frequently corrupt the indexing, making the edited entries (and sometimes entire pages of entries) completely disappear. There are several Community threads on this topic, all of which now appear to be closed. Was there an update on FamilySearch on this topic?
There was an assurance that this issue was being fixed, and that all of the corrupted entries/pages would be restored. I saved links to dozens of census pages that I inadvertently corrupted due to the Index Editor bugs. I just went through and re-checked them, and I would note the following:
- The Index Editor seems much less likely to corrupt things now, but I just had another occurrence, so this is still not fixed: The MacIver/Nuckols family on this page (https://www.familysearch.org/ark:/61903/3:1:S3HY-6S69-M5M) was improperly grouped with the Logue family on the next page. When I tried to put them into a new record group, the MacIver/Nuckols entries disappeared again, and the Logue entries now appear on both pages.
- A small number of the entries that disappeared during my previous edits seem to have re-appeared. However, the missing entries on the vast majority of these corrupted census pages are still gone. I wonder if somebody just manually fixed a couple of the pages that I previously reported, but I did not share the longer list of corrupted pages—and perhaps that indexing cannot be automatically restored?
Answers
-
@SteveLinke We have documented (and engineers are aware) that using the drag names tool to group/create a family puts a lock on the entire family as soon as the process is completed. To the user it may appear that all is well until the page is refreshed and the names are not seen in the index. Nothing has been corrupted—permissions to view have been removed and generally the approval process is taking from one to two weeks and then the names are again viewable in the index. This is a bug and engineers are working to resolve it ASAP.
There is at least one other smaller issue with editing. When adding a new event to one persona, if a second one is added to a different persona without refreshing the page the event will be attached to the first persona.
1 -
@SerraNola Thanks for your reply. Based on my own observations, the disappearance of entries after re-grouping is most common when families span two pages of a census, although I think it happened/happens occasionally even when the entire group was on one page. Hopefully, the MacIvers/Nuckols will reappear in the next week or two.
That said, my use of the Index Editor a year and more ago caused the disappearance of many census entries. The first few dozen times this happened, I thought I was just doing something wrong to view them again, and I did not track the pages that I corrupted through the use of the Index Editor. I then checked this Community and saw that many other people were experiencing the same issue.
We heard that nothing was actually being lost, and that all of the indexing would get restored once the bugs were worked out. Even so, I tried to save a link to all of pages that I was corrupting from that point on. I just checked, and these pages that I corrupted (many months to over a year ago) are still missing the indexing.
The problem is that, the people in those entries no longer even show up in searches, so people trying to find their ancestors will not find them. And, if people had already added the entries to sources for their ancestors, the links to those source documents just return errors, as if they do not exist.
Here are links to 8 examples of pages that I corrupted using the Index Editor (I have many more):
1870 Census (MA): Household 49 (Fee)
https://www.familysearch.org/ark:/61903/3:1:33S7-9YBJ-2PW?view=index&action=view&lang=en
1870 Census (NJ): Lines 35-40 (Dunham)
https://www.familysearch.org/ark:/61903/3:1:S3HY-DTJ3-SZP?view=index&action=view&cc=1438024&lang=en
1895 Census (MN): Henke household
https://www.familysearch.org/ark:/61903/3:1:S3HY-6X8W-P7Y?view=index&action=view&lang=en
1895 Census (MN): Household 94 (Henke)
https://www.familysearch.org/ark:/61903/3:1:S3HY-6X8W-PML?view=index&action=view&cc=1503031&lang=en
1930 Census (IL): Lines 42-44 (last three on page)
https://www.familysearch.org/ark:/61903/3:1:33S7-9R4V-CYM?view=index&action=view&lang=en
1950 Census (CA): Lines 2-5
https://www.familysearch.org/ark:/61903/3:1:3QHJ-5QHW-M7YW?view=index&action=view&lang=en
1950 Census (IA): Lines 2-30 (most of page)
https://www.familysearch.org/ark:/61903/3:1:3QHN-GQHW-TKLZ?view=index&action=view&cc=4464515&lang=en
1950 Census (MN): Lines 1-6 (Rabenhorst)
https://www.familysearch.org/ark:/61903/3:1:3QHN-GQHW-KX3C?view=index&action=view&lang=en
1 -
In testing over a dozen grouping of families, the index disappeared every time the process of dragging a persona to the New Record Group box was completed. Dragging additional personas to the new group also removed their "permission to view". I did not try it with groups spanning two pages but I assume it would be the same outcome. Out of the testing group, two indexes did not return in the two week time period.
You have done an excellent job defining a serious issue with Search. You are correct that the loss of indexing that occurred with earlier editing has not been restored, even though the "bug" was fixed. You are also correct that when a persona is "locked" it is not search retrievable. Here is what the index detail looks like for one of the missing indexes in your examples:
You can see that Donald Alexander and his family are all still indexed but the padlock tells you the data has been restricted from view. Obviously, this is not a normal restriction stemming from the Contract and Compliance team.
We are getting many reports of people not indexed and the user is unable to index them. I believe most (if not all) of these—when there is one family or a few people missing, are a result of previous indexing or recent family grouping. They cannot be re-indexed while still locked.
You mentioned in your initial post that sometimes whole pages are missing. That might be partly the same issue but we also have some collections that are missing large portions of their original indexing from a different cause. In talking with engineers, we are told it is a multi-faceted issue that is being worked on one piece at a time. Be assured that they rank this problem as an extremely high priority, but one that will not be resolved overnight.
0 -
@SerraNola This one has been locked ever since I made a tiny indexing change in, iirc, 2023 (correcting him from Thomas I to Thomas J): https://www.familysearch.org/ark:/61903/1:1:XH2V-K3Y (and there are surely no groupings anyone could do anything to, on the 1810 census).
0 -
Maybe I wasn't clear in my post. The old editing bug has been fixed, but lost indexes have not been restored. The bug that has not been fixed and is still wreaking havoc is the one with grouping using the New Record Group. I would advise everyone to stop using this tool until we get an update.
1 -
Ah ok thanks, that makes sense.
Could you not get the engineers to disable the havoc-wreaking grouping tool until they have fixed it, given the pretty small percentage of users who are likely to see advice posted here?
0
