I have a database in RootsMagic 10 that I previously imported from Gramps. Now, I need to update the data in the same way, meaning I will perform another export-import. My question is, what should I do now (or at least during the next iteration of database updates) so that RootsMagic can clearly identify the same individuals? Would the FamilySearch ID help with this? Or maybe some other identifier?
Gramps does not add unique IDs to new persons, so there’s not much to merge on. Automatic merges by the persons’ data may work, but I never tested that, because I only have RM Essentials here.
This means that for practical purposes, you can best select either as your main tree, and let the other follow that, by creating a new tree from the GEDCOM export of the main tree.
The FS ID could help if you run into problems. However the ‘fix’ might need to be done either the Sqlite and/or some manual update – myabe someone can advise – right now I would have to test that but I do not have time for that, I have not played withe Gramps database directly but you might be able to direct database to database transfer if built correctly.
I don’t really have enough information about how you are using Gramps and RM together to make a reasoned recommendation. For example, can you just delete your RM database and import again from Gramps into RM? I suspect the answer is no, but I don’t have enough information to know for sure.
The latter depends on your workflow. In mine, Gramps is the mother of all trees, and RM is one of the prograns that I use as an add-on to communicate with Ancestry and FamilySearch. I import any data that I think is worth it back into Gramps, so that the data in RM, and other programs, can always be refreshed via GEDCOMs, from Gramps.
Things are not always easy though, because GEDCOM dialects can get in the way, so we can’t simply use WinMerge to figure out where the new data is, in the GEDCOM exported by RM.
Probably it can help in merge duplicates tool when FamilySearch ID is used as all other user data during duplicates searching. But it doesnt help in the Compare tool. I tested it when fully changed person with the same FamilySearch ID is compared between two DBs. RM doesnt identify them as the same person. But _UID worked good for this purpose.
So, I generated _UIDs in Gramps for all people - used own python script for this. Then export-import to RM. Then I made some changes in Gramps and imported the new gedcom into RM as a new database. Then used Compare tool. And this tool really used _UID for comparing - it works very good with 100% result. But then I modified some people in RM end exported to gedcom. I see that _UIDs of changed people isnt added to gedcom. As I understand, my generated UIDs were not fully valid, looks like I need change generator algorithms, but I didnt test it yet.
So, I would like generate _UIDs for RM in Gramps (my main DB is in Gramps). I can do this automatically when create new people in Gramps. And then will export gedcoms to RM and compare the by _UID. I need RM to push people to Familysearch. Such is idea.
But I also need have the same UIDs for people who received Familysearch ID to import new Familysearch ID backward in Gramps.
That’s right, and it works nice with Ancestry too, because RM can download your tree from there with any pictures that you attached there, which is especially useful for scans attached to sources, that you don’t find in GEDCOM files exported by Ancestry itself. Ancestry GEDCOMs also have other quirks, like full month names in dates, and UIDs exported as UID i.s.o. _UID, and RM corrects both.