Best way to integrate foreign gedcom?

[semperor]

Hello².

Please help me to find the best way to solve the following task with GRAMPS.

I have a family tree with about 2000 people. I got a GEDCOM from a researcher colleague, who also researched about 400 of these people and people unknown to me. I now want to integrate his results into my family tree. By hand, i.e. looking at each person, family, etc. individually and transferring the additional data and sources, I would probably need months to years.

How would you do that? How can GRAMPS support me?

Thank you.
Hagen from Germany

[prculley]

Integrating a foreign GEDCOM file is one of the more difficult problems any genealogy program faces. Gramps is no different.

I would start by opening a new tree and importing the GEDCOM and looking it over for issues. If there are people that you don’t want, this is a good time to delete them. If you trust your own work for common people, you may want to delete the common people, at least those away from edges that may have new and interesting people. You don’t have to do this, but it does get you some familiarity with what is likely to need further work and will save you some effort later.

If you do this step, export your work as a Gramps XML file.

Backup your own tree in case you decide you need to start over.

Once you are ready to start integrating, import the foreign file (either the GEDCOM or XML, if you did that) into your own tree. Everyone and all the data is likely to be imported as new objects, even if they are common between the two trees.

Now comes the hard part, merging. It is largely a manual process, usually starting with people. One tool which can help a bit is the “Tools/Family Tree Processing/Find possible duplicate people”. This tool may help with matching up common people and performing the merge. Or you can just sort the persons by name and do it manually. After a merge, it is a good idea to open the person with the person editor and look at events, citations, etc. and see if duplicated data is there, and delete it or make a note to merge the item if there is good stuff in both almost alike objects. You can even go to the view containing the objects and perform the merge while you have a person edit open.

Then you can do the families, no tool for this.

If you delete objects like events, notes, media etc. from your families or persons along the way, you will end up with these ‘unused’ objects in the other views. You can use the “Tools/Family Tree Repair/Remove Unused Objects” to clean these out of your tree. You may have to run this multiple times to get them all, as removing an event may make a place unused etc.

As you can see you will end up doing a LOT of manual work, no real help for this at this time. Performing a merge automatically is going to require almost Artificial intelligence, and we are not there yet.

Make frequent backups, in case you need to blow off a big change. And remember, you can Undo changes if you make an error within the same Gramps session.

[semperor]

Wow. Many thanks for the detailed answer!

This sounds to me like a very well thought-out approach. Is there a best practices section somewhere where you can store this?

[PLegoux]

@semperor

Is there a best practices section somewhere where you can store this?

I don’t know if there some section but you can store this answer in your own section using this symbol:

You’ll find it after in your profile section, here:

[DaveSch]

Paul’s response has been tagged with a green check mark and Solution. I am sure Nick Hall made the tag as he is the administrator.

Hopefully the next user wanting to know about merging will find this thread and see the post designated as a solution. Not sure if there is a special grouping for these posts making them even easier to find.

[emyoulation]

The Help Category is one of the sections where the Solution flag option is enabled. It allows the person who posted the question to mark their preferred solution. And it also appends a linked ‘solution’ blurb at the bottom of the original posting. (So the posted answer stays in its threaded chronological order but also has a shortcut added at the top of the thread.)

This thread has also been linked from the “See Also” section of Gramps’ general GEDCOM wikipage.

[TerryS]

I understand that many of you know each other on a first name basis but as a newbie to the forum I find that exposing real respondent and admin names in a public forum, where many more are expected to arrive, somewhat unusual. Personally, On a personal level I like it. At a system or wiki source level it probably needs filtering - over to the admin for a policy decision.

[Nick-Hall]

@TerryS

over to the admin for a policy decision

It’s probably better to use usernames rather than real names. The avatar tooltip displays the real name, but it is not immediately obvious who the poster is talking about.