How Compare Myheritage GEDCOM with gramps db?

I have a GEDCOM export from myheritage. The people are written in Russian there. But my local Gramps DB is Ukrainian. So, I think I can not use any matches, comparing, … of Gramps or any other program to merge the data. A lot of people are already in my db, a lot are not in the db yet. I have several questions:

  1. what is the best solution for today to merge myheritage to gramps? Gramps doesn’t have any myheritage ids and vice versa.
  2. I would like generate my gedcom exports not only in Ukrainian but in English also. As I know Gramps doesn’t support names localisation. Have you any ideas/solutions for localisation? What I think, it could be a separate script out of Gramps which could translate GEDCOM file (names inside the file). Maybe you have better ideas?
  3. I think, I can not simply merge GEDCOM into my db, but maybe I could mark all the imported people with any tag, then merge 2 dbs into one. As a result I will receive one big db, some people have the tag, and another don’t have. Then I run any script which comparing tagged people with untagged. I even can define starting the same duplicated people in each of the group. And then the script will only count their children, brothers, parents, etc. If counts are different, it means, I need manually add some family members to my untagged tree. So, the script will work as a checker, which compares not names, places, but it compares only counts, how many people each family has. What do you think about such idea? Maybe anybody has better solutions…
    I would like hear your thoughts. Thank you!

As for my first question I asked something like this Importing and merging with MyHeritage which posted in 2022.
What is the best working solution for today in Gramps?

I have been modifying the importmerge plugin for my needs as I want to import from MyHeritage, Ancestry and another db my aunt aus worked on years ago. The plugin begins working for me. The code is messy and not fully functional so I hesitate to share it.

1 Like

@Nick-Hall ,

Do you recall who has been in discussions here or GitHub about UIDs/UUIDs ?

It would be good if people could see where external IDs imported (from Gramps XML, Gramps Web Sync, GEDCOM, Ancestry, FamilySearch, FindAGrave, WikiTree, et cetera) are being being stored. And how they can be preserved and leveraged as a hint to help merging.

1 Like

I don’t think that the situation has changed, since then. I have a tool that adds _UID attributes to my own database, but that tool was rejected, because part of the code came from ChatGPT. I already had many persons with _UIDs, because I migrated from PAF, which once set the standard for _UID in GEDCOM files.

There is a setting in Gramps to add a standard source during GEDCOM import, which you can switch on before you import the MyHeritage GEDCOM, and switch off afterwards. You can then filter on that. It’s not a tag, but you can use it in a person filter, anyway, I guess.

I know from the MyHeritage Family Tree Builder that it can work with languages for person data, meaning that when I use it in Dutch, I get empty names when it syncs data in English. Is that what you mean? Or does the difference between Russian and Ukranian mean that most names have small differences in spelling, so that a direct comparison fails?

Did you create that MyHeritage tree from scratch? Or was it an upload of a Gramps GEDCOM that you expanded on-line, with all sorts of matches? And do you, or can you, consider one of these trees as your main tree?

In this situation, I think that I would import both trees into a single database in RootsMagic, and let that search for duplicate persons. It does that much faster than Gramps, and is also a bit more efficient, because it will not create duplicate events when they have the same type, place, and date. It has an option to find names that sound alike, which is slower, but still way faster than Gramps. RootsMagic Essentials is free, and there are versions for Windows and MacOS.

How many persons do you have in these trees?

I didnt know thanks for that!

Probably yes. I mean that each name can be translated to multiple languages. And I can choose a language when export to Gedcom. If a person isnt translated, I probably receive a person with name like this “Has No Translation”. But I understand, Gramps doesnt support multilanguages for names. Another way, I think, It could be any script which opens gedcom and replaces all strings to another. It could be alternate and maybe more easy way to receive what I want.

I made gedcom export from Gramps to myheritage (1500 records). Another guy found mathes and imported them into his own myheritage tree. Now he has 4000 records. And these 4000 records he exported into gedcom for me. So, 1500 records are the same as my local in Ukrainian, another 2500 records are Russian.
Another one… His records have middlenames in the firstname field, but my records have middlenames as surnames with “patronymic” type. So, I have:

  • different languages
  • different way of middlename saving
  • my 1500 records dont have myheritage UIDs yet
    Looks like the better way is dont merge merge them into one DB because I will receive more unresolved duplicates, several languages, different locations of middlenames :exploding_head:

I still can not decide what is better:

  • save middlenames near firstnames
  • save them as now in surnames

Both ways looks poor for me. I think I hope maybe some another special reserved fields will appear for middlenames )). Maybe additional reserved types for middlenames could solve this problem for all users who works with middlenames. In this case middlenames could be exported in Gedcom in more valid way, and they also could be used in person name formatting in the settings.
And also I dont understand why first name has one field only, but at the same time I can add multiple surnames? I have real cases when I need add multiple firstnames like it works for surnames.

These are my biggest headaches :sweat_smile:

The English version of this is in reality there is no such thing as a
middle name you can either have multiple words (mine max out at 4) in
the first name field and then a surname which can also be more than one
word sometimes separated with “-”.
So to me it is never an issue because the “middle” name is a concept
that just does not exist.
How many fields would you want for “given” names and how many for
“family” names?
Create you own convention for the fields that are there.
phil

Any my own conventions leading to exporting and reporting problems. And also names formatting can not take into account middlenames which I would like see right after first name but not anywhere else…

Pls, dont say that middlename doesnt exist. I can understand that it doesnt exist in Gramps, it doesnt exist in gedcom, but it exists for many millions people in the World. And some of them have daily troubles with it. 50% troubles due to Cyrillic, another 50% doe to middlenames. Unfortunatelly. Its not about Gramps, its about any genealogic software.

We have the same here, where my middle name, Frank, is just that, sort of, although it is often used to honor an ancestor. And in Germany, it is sometimes the call name.

In Russia however, it’s often patronymic, like for Boris Nikolayevich Yeltsin, who was a son of Nikolai, according to Wikipedia. And in such a situation, it’s not the same as a 2nd given name, like mine.

1 Like

Note I started my previous post with the word “English” which is of course the main language or basis for the majority of Genealogy Software.

And I also pointed out “for me” when referring to “middle name”
Assuming that most people would realise that is where GRAMPS is coming from I was not denying that there are other naming conventions.

However anything requiring import/export via GEDCOM is going to suffer from the above.
phil

1 Like

Even though Gramps was created by someone native to my home region (western Pennsylvania, USA) and it suits my needs, I am intrigued by the features to make it work with other cultural biases.

The date formats, name guessing and language customizing are good starting points. But they need “fleshing out” by developers whose cultural bias is not well supported.

But to facilitate that, switching the bias needs to become less painful… Dates format switching should not require re‐starting Gramps. Adding another language should not require re-installing Gramps. Name guessing should also guess the name type upon which the guess was made. (“Patrilineal” is the default name guessing bias. And “Taken” should be the default for a Married Name surname that differs from the “Patrilineal” one but that matches the spouse surname.)

1 Like

Agreed.
Which is why I stick with the Scottish of all women retaining their
maiden name to the grave with largely this being the surname on any
females gravestone in Scotland.
None of this Joan (maiden name) (married name) or
Joan (married name) born (maiden name).

Well, although Gramps is ‘made in America’, it works quite good in western Europe, because I can use prefix fields for ‘van’ or ‘van der’, which is the Dutch way to spell names like Vanderbilt, which was derived from the Dutch ‘van der Bilt’, which would be indexed as ‘Bilt, van der’ in the phone book. Gramps also supports the German call name, but not the Dutch one.

These things can not be found in most other American software, like Ancestral Quest, or RootsMagic, and because of that, we’re used to expect more from Gramps than from the others.

Moreover, when you look at GEDCOM 7, there is a movement to accept names as a sequence of parts, which can have their own type, so that it’s finally possible to support cultures like Vietnamese, where the surname comes first, and the ‘first name’ comes last, with some sort of (family) nick name in between, when I recall that right.

2 Likes

I must add that I gave up on MH and FTB a few weeks ago, because MG refused to react to my bug reports. That got me so mad, that I cancelled a running membership, without the right for a refund, simply because the mental costs were bigger than the Euros that I paid for that year.

What I do remember though, is that when you have two languages in FTB, and maybe even more, that you have the chance, if you know how, to copy all missing names from one language to the other. I tried that with Dutch and German, and it worked, but I found it quite silly, because they could also have chosen to export the name in the other language, when the preferred one was not available. That would be a big time saver for the help desk, although I bet that their bot wouldn’t mind. And that was so dumb, that I complained about it on Trustpilot, which immediately resulted in me getting a special email address for their VIP support. It didn’t help much though, because they didn’t really cure anything. The only things that I got were friendly words from a human, instead of the bot.

Long story short: If you use this trick to copy Russian names to missing fields in the Ukranian version, you will have a GEDCOM that looks better, but it still has the ‘middle names’ a.k.a. patronynics in the wrong places, and you still don’t have the IDs to match on. And that makes me think, that you better do what I do, and that is to select the branches that you really like to add to end-of-line persons in your own tree, export those to new empty trees, correct the names, and then import those to your main tree. And maybe you can use some Isotammi tools for those ‘middle names’ too.

And many of my fellow countrymen, and -women, would just tell you to get rid of MH ASAP.

1 Like

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.