Download tree through 4th cousins from FamilySearch

The short version of my question is: Is it possible to download my family tree through my fourth cousins from FamilySearch, and, if so, what is the best practice to accomplish this?

What I have tried

I first tried an import starting with my FamilySearch ID (FSID). I suppose it should have been obvious that that would not work, because it only captures ancestors.

I next recorded the FSIDs of my 16 great-great-great-grandfathers and tried to serially import them from FS into a RootsMagic database. However, links of the people between the “sub-trees” were not created.

Now, I am downloading all of the 16 sub-trees separately and am combining them into a separate “master” database file through drag-and-drop (everyone) followed by AutoMerges.

The last strategy seems to be working, but I am wondering what might be getting lost: Some links between people? Multiple marriages? Other?

Also, is there a better strategy?

Why I am doing this

My mother created a huge genealogy database over a 40-year period–most recently in Family Tree Maker. She has a little over 10,000 people in there. She has worsening dementia and wanted me to take over. Unfortunately, she was not a professional and did not record or maintain any sources.

After considering many options, I decided to try to add her information to the FamilySearch tree and attach as many source documents as I can. I have done about 3,000 people so far, and I have tried to clean up messes on the FamilySearch tree as I have gone along (and tried not to create too many of my own).

I realize there are risks of using FamilySearch as the main platform, but I am hoping the benefits outweigh those, and I have found it to be pretty powerful and amazing for a browser-based application–particularly in combination with the search links. I am so used to the way it works that it can be frustrating trying to use desktop software, including RootsMagic.

In any event, I thought I would periodically do this download through my fourth cousins to get a “snapshot”/backup of what is currently on FamilySearch. It is looking like that will include around 8,000 individuals. That is a large number, but going back another generation would have been completely unmanageable, so I settled on that level.

I am also wondering whether I can use my combined master file in RootsMagic to monitor changes that occur over time on FamilySearch, or whether it is truly just going to be a snapshot. For example, if new children get added on FamilySearch, or if IDs get merged or deleted, I am not sure what happens when I open my file in RootsMagic. I don’t think it is really that “in sync” with FamilySearch to reflect such changes.

Is there something useful that can be done to track changes and keep such a local database updated?

I would also note that I frequently get “Unexpected Errors” in RootsMagic 8. For example, when I do drag-and-drops of newly imported “sub-trees” into my “master” database, I can’t do two in a row without getting a “Cannot perform this operation on a closed statement” error. If I close all files and re-open them, it then works fine.

Also, when I try to do an FS import from a 3x great-grandparent (e.g., 5 descendant generations), it will often fail the first time and only import the first person. But then it will work on the second attempt and import the hundreds of expected people.

And, as part of my process of entering information on FamilySearch, I first imported my mother’s Family Tree Maker file into RootsMagic 8, and I am adding a color code to each person when I finish them, so that I know what I have completed. Over the course of a few hours of just adding the color coding, the program will start complaining about errors, which persist until I close and re-open the database.

This post was flagged by the community and is temporarily hidden.

Hello Steve,

Let me start with a warm welcome from The Netherlands. I’m using Gramps as my main program, because RootsMagic doesn´t speak Dutch, but I use RM 8 a lot to exchange data with FamilySearch, and download snapshots from Ancestry.

Importing a tree up to your 4th cousins is easy, and shouldn´t take too much time, if you set the right numbers in the Import dialog, meaning that you set the number of generations to 5 for both ancestors and descendants. And I write 5, because 1st cousins means 2 generations, so 4th cousins means 5. I just tried this here, and the total import gave me a few hundred people, not thousands. You may see more, when your cousins had a lot of kids, and the whole import can be done in one go.

Once this is done, you can use FamilySearch Central in the Publish tab to let RM 8 look for changes, and initially, you may be surprised by the fact that the program will tell you that it has new information for most of them. It looks weird, because this happens right after import, even if no-one else made a change on either of these, and the reason is that the import is not complete. It will not process alternate names, and non-vital events, and it does not import sources either, so every person that has that extra information will be shown as having new data. And I think that you will also see people marked as such when they have relatives that were outside the scope of your import.

FamilySearch Central has a handful of filtering options, so there is no need to use color coding to check which people have changes, or haven´t been matched with FamilySearch. RM 8 can do all those checks for you, if you want.

If you have the full version of RM, you can keep a single database in sync with both Ancestry and FamilySearch, so there is no need to work with multiple files whatsoever.

Note that, like Rooty, I would also recommend that you focus on the Ancestry tree, and not on the one on FamilySearch, because the latter can get quite messy, and your mother’s work is already on Ancestry, or isn’t it?

If it is, you could of course stick with FTM too, but if you like FamilySearch too, RM is the best desktop program to work with both sites, and get hints from Find My Past and My Heritage, which can get very handy when you have relatives in Europe. And if your mother’s work is on Ancestry, I’d start downloading that tree, and then use FamilySearch Central to check for added value on FamilySearch, and to publish where you see fit.

I do see some glitches sometimes, especially when I try to import new branches, but most of these can be cured by closing and opening the database, and using Duplicate Search and Merge to reconnect people that got isolated somehow.

I hope this helps a bit. :slight_smile:


1 Like

Hi Enno,

1. Greetings

Thanks so much for your reply from The Netherlands. I tried Gramps first, but I found it too complicated, and then I decided I wanted to interface directly with FS and Ancestry, so RM was the clear choice. Honestly, I have become so accustomed to entering data and attaching sources directly on FS, that I actually prefer that interface (if only it wasn’t subject to getting so messy).

2. FS Import

In my experience, it is not easy to import a full tree through 4th cousins from FS using RM. If I am making things more complicated than necessary, let me know. As I understand it, when you designate the number of generations of ancestors in the RM import, it only imports the direct ancestor, their spouse, and children. It does not continue down to capture all of their descendants. I just tried that again, designating 5 generations of both ancestors and descendants, and that seems to be the case.

In order to capture all of my 4th cousins, I had to separately import 16 sub-trees starting with each of my 3x great-grandfathers (I imported a total of 13,471 in the 16 sub-trees). By capturing each sub-tree separately, all of the relationships were also captured. If I sequentially imported from the 16 3x great-grandfathers into the same file, the overlapping relationships were not linked.

After creating the 16 full sub-trees, I sequentially combined and AutoMerged them into a master tree. The AutoMerge identified the vast majority of the duplicates (5,656 in my case). I then did a Merge Duplicates, which identified 167 additional duplicates, and, finally, I manually merged 99 remaining duplicates. That left me with 7,549 total unique FSIDs out to my 4th cousins (if I am understanding things correctly).

FEATURE REQUEST: It would be nice if the FS import could be programmed to do this automatically (download out to X number of cousins). If that is not practical, though, it would at least be nice to be able to designate an automerge using the FSIDs. Those are unique numbers for each person, so it would be logical to automatically merge them.

There is a selection in the Duplicate Search function called “Find people with the same reference numbers,” which I thought would do this, but it did not. Perhaps there is a way to automatically copy everybody’s FSID into their “Reference No” fact field, but I did not try that.

3. FS Central - getting additional data

So, does each separate source and fact from each separate person have to be imported separately into RM? If so, then perhaps it is not practical to use RM in the way I was hoping. I could use RM to maintain the basic tree with just the vital information, but I have probably effectively already committed to using FS as my main platform and should just use that (hoping that nobody else will mess up any of the work I have done).

4. Color coding to track progress

I am only using the color coding to track my progress on entering the data from my mother’s FTM tree into FS. I was given an FTM file that had a little over 10,000 people in it with a mixture of birth, marriage, death, burial, and other data, along with some narratives. Frustratingly, none of this data had ANY SOURCES attached to it.

When I looked on FS, a lot of the people were already there–with sources. I bought RM 8 and imported my mother’s FTM file into it, and I have been going through each person, one-by-one, adding unique information to FS and attaching sources that are found. I then mark each completed person with a color. When I am done with this process, and everybody is marked, I intend to “retire” that RM file and was hoping that I could make a new, complete RM file through a FS import. Based on the apparent necessity to import every source and non-vital fact separately into RM, perhaps I will have to abandon that part of the plan and just use FS going forward–which is OK.

5. Ancestry vs. FS

I chose to start with FS, because a bunch of the people were already in the shared tree with sources attached, and I wanted to add my mother’s work to this collaborative project–and because it will always be free and open. I assumed that it would have been even more work to attach all sources to 10,000+ people uploaded to Ancestry. I am still not clear on what would have been the best choice, but I think I am pretty much committed to FS at this point.

I am very interested in your comment that I “…can keep a single database in sync with both Ancestry and FamilySearch.” It seems that the syncing would only include the basic vital information–and not all of the extra information or sources. Is that the case? I am new to all of this, so I am not sure about all of the best practices.


Hello Steve,

Oops, sorry about my misunderstanding. I ran another check on my test file, and it seems that I got fooled by my programmer’s mind, that thought that the process would be recursive. It’s not, so you’re right, and I think that I got fooled by the fact that the import includes uncles and great uncles, but I now see that they’re only included when they are ancestors’ siblings. And that’s something you can’t switch off, so it does slow down import if you want ancestors only. There is another program that can exclude siblings, and it would be on my wish list for RM too.

I am not sure why sources and non-vital data are not included, but my guess is that FamilySearch wouldn’t like that kind of bulk downloading, for a couple of reasons, including speed. I know that RM’s competitors don’t import these either, so you’d have to import these one at a time.

As far as I can check, this applies to syncing too, in the sense that FamilySearch Central will only show you which persons are different, and lets you mark the items to be downloaded or uploaded, one at a time. And the latter is very important for FamilySearch, because it prevents bulk uploads, which might create a big mess.

Since I’m only using RM Essentials, I can’t comment on the syncing with Ancestry, but I think that it’s more automatic. And I think that, because it makes sense since you’re always working on a personal tree there, so you want that to be in sync with your local copy, and it wouldn’t be good if FTM can do that, and RM can not. It’s not that import for me, because I only use my Ancestry tree as cousin bait, and do all merging of new data in Gramps.

I’m a freerider on Ancestry too, btw, because most of my tree is European, and most if not all local sources are free.

Do you have a subscription on Ancestry?



Hello Steve,

I did some further research, and there is a way to download your ancestors and their descendants, with sources, notes, and everything, with a Python program called getmyancestors. I remembered that it was mentioned on the Gramps forum, and tested it with 5 generations up and down, and it gave me a GEDCOM file with more than 2000 people, and almost 8000 sources, in about 10 minutes.

You can find more information, and source code, on GitHub:

I have tested this in Linux Mint, with Gramps, and RM 8 (running in Wine), and the GEDCOM import works fine in RM 8.

I have no idea how easy this is in Windows, but Python 3 is available for that platform, and I just found a long thread about running it in Windows on our own forum, which uses the same software as this one.

I haven’t tried anything of this, so everything you try is at your own risk.

It is download only, for obvious reasons, but I do hope it’s a good start anyway.



Thanks, Enno! I will check that out.

OK, please report back, if you can. I tried a quick test on Windows, but I saw some errors here.

I was mostly able to make the getmyancestors Python script work.

One issue was that I found that I had to install Python 3.9, because the newest version of Python (3.10) resulted in the following error: [AttributeError: module ‘collections’ has no attribute ‘Mapping’].

After installing Python 3.9, I ran the following from a CMD window: pip install getmyancestors

I then launched the GUI version of getmyancestors by entering the following in the CMD window: fstogedcom

I was then able to enter my FS username and password, and an input window came up that is very similar to the RootsMagic FS import window. It seems to download all of the sources and additional facts that RootsMagic does not–and it does it relatively quickly.

The downloaded GEDCOM file was easy to import into RootsMagic, and the people seemed to sync with FS. However, not surprisingly, the sources and additional facts are not “linked” to the equivalent information on FS.

Also, I think I would still need to do separate downloads for all 16 3x-great-grandfathers, and then try to merge all of the overlaps, as I did previously in RootsMagic. I was not able to figure out a way to download both ancestors and their descendants in a single run.

So, the bottom line is that the getmyancestors tool can provide a much more complete snapshot of a tree on FS, but there really isn’t a connection made that would allow future syncing.

FEATURE REQUEST: I can understand that allowing bulk uploads of huge numbers of sources and additional facts from RM to FS could be very disruptive to FS. However, it seems like RM could be allowed to download those things from FS and create the links necessary to upload future individual changes.

My situation may be unique, but it might be nice for people who are just starting on their genealogies–and who are willing to take the risk and deal with some messiness on FS–to do an initial download of their trees as they exist on FS (including all of information available there). Then, they could work to add their new information one fact/source at at time for upload.

Well, that part is why I mentioned this. You can do this by putting 5 in both generation fields of the fstogedcom dialog, and I’ve seen that work on my tree. That’s how I got those 2000 persons in one go. That second 5 is exactly what I meant in my 1st message, not realizing that in RM it counts from the start person. In getmyancestors, it’s the number of generations down from the ancestors, so if you specify the same number both up and down, you’ll get exactly what you asked for. It says so in the readme too.

Further on, I don’t understand what you mean when you say that there is no connections made. As far as I can check it is, because the program puts the FS ID in the GEDCOM, and RootsMagic reads that. In my test, all persons that I imported were immediately marked as linked.

I understand the problem with sources, but what you get here is far better than what you can get with any commercial program.

Thanks for trying Python 3.9. I gave up, because I didn’t want to invest much time in Windows here.