Return of GEDCOM Import

Started by Mike Stangel on Friday, February 22, 2019
Problem with this page?

Participants:

Profiles Mentioned:

Related Projects:

Showing 91-120 of 191 posts

It is not possible to revert an upload to my knowledge. Every single profile has to be changed one by one and every double profile has to be merged with all the data conflicts.

Please ask Per Anders who is doing a great job.

Per Anders (PA) Johannssen

Two of the metioned users are informed - Molly will be informed today.

Thanks Remi-

Remi Trygve, can everybody reverse an upload? I guess not.

Who can do it? The uploader (of course) and a curator?

(No Name) - only the uploader, or Geni staff if necessary.

(No Name) - in addition, had you seen Mike's post on the previous page: https://www.geni.com/discussions/192952?msg=1300115 -- that is another way possibly to cut down on the work - tho not being a Curator, neither you nor I could mark it fictional on our own - but we could cut and send the link and info to Mike, or cut and ask a Curator to mark Fictional and let us know whether they will send Mike the link or if we should

What do you mean common good? I hope you don't mean merging duplicates.

Thanks guys, we try to make the best of your information.

Duplicates need to be merged

Kenneth, we know that duplicates have to be merged and data conflicts or other conflicts have to be resolved.

However more or less new users who think that they upload the "latest and greatest" do not know that.

I have forwarded the message from a Pro user directed to the Geni Staff so that the Geni staff would see it. If you are a Pro User please create a ticket if a GEDCOM import goes wrong.

The hidden message has also personal information about a user that did very likely a very very very large GEDCOM import, that did go very wrong. Mike Stangel your reaction please and please repost the message without the personal information and your reaction.

Are there limitations in the current gedcom import, like a maximum number of generations and/or max number of profiles?

(No Name) and Private User - Importing a GEDCOM does not directly allow a user to change existing profiles. Of course when duplicates are merged the effect may be the same, however the extend of such damage should be limited for two reasons:

1. When merging new profiles imported / managed by a single user with older profiles (especially if they have multiple managers), the older / original data (including names) should be preserved, and the imported names should go into the data conflict table.

2. The importer should not allow more than 3 generations of profiles to be imported, when there are nearby tree matches. That should be worst-case, and often it will be less than that.

Now obviously Gerhard wouldn't be posting if these damage-mitigation controls were working well, so I think I need to better understand (or maybe see examples of) the cases where this user changed names of many well known and well documented and accepted names of Swedish nobilities and other Swedish people. Can you send me a few examples, or post your understanding of how the controls above did not prevent it?

Kenneth Ekman An import must start on a profile born after 1800. It will import 5 generations at first, before stopping to look for matches. If matches are found on a particular branch, the importer will not continue on that branch. Instead, the users must resolve the match (accept / merge or reject) and once a profile is merged into one with parents, the importer will not continue importing up.

If a branch has no matches, the importer will continue importing 3 generations at a time, looking for matches. Once those 3-generation imports get back to pre-1600, even if no matches are found it will not continue with another 3 generations of ancestors. (Meaning it's possible the import leaves off at 1601, continues to import 3 generations above that, but then it will not do another 3-generation ancestor import after that).

Merging these GEDCOM duplicate imports is now a full-time job!

I'm importing a large file right now, 70.000+ profiles.
First, does that mean that everything before 1600 will not be imported at all?

The merging-process is doing fine. If you look at the progress daily, and do the merges and fix the mistakes, then no problems at all. But, don't do an import and leave it at that!
In my opinion - uploading a large GEDCOM requires responsibility.

These are the problems that I encounter:
- The gedcom page is most of the time not available
- Baptism dates are not imported
- Parents without data, of people with data are not set to deceased and/or set to private. Say, a person born in 1897, you know he/she is dead, is not said to deceased. And the parents (because she is in-law from your tree), without data are also imported as being alive.
- Some people, without data, but with direct family living before 1800 are also private and alive.
- Merging private profiles, especially people who are alive today, you need a curator to do the merge request.

But I love the function. When you have so much gathered data this is really helping.

Is any one of the 70 thousand profiles not already on Geni?

Yes, they are. In fact, there was already a profile for me as well.
I think 25% to 30%. It is great and satisfying to be able to connect some lose threads already on Geni by my import.
The merge center is ideal to tell me if there are similar profiles. And in 95% it is correct. I perform the merges, solve data conflicts and voila, on we go.
I correct a lot of profiles already here on Geni on the way.

I am and another dutch Curator are helping Maarten Alexander Brandts Buijs and if needed we will disconnect the double profiles. But so far it is a lot of work, but it is going good.
Still 5% of 70.000 is a lot of profiles. But I have something to do.

Within those 5 percent are also siblings with the same name. Those I disconnect myself since I learned the trick. A big thanks for those two curators, they were helping me a lot, especially those first couple of weeks.

What would happen if you try to merge them?

Is it obvious to you that these are five sets of the same couple?
Daughter of Hans Jensen Åbenterp d.æ; Hans Pedersen; Hans Larsen; Hans Hansen; Hans Jensen; Margreth Pedersdatter; Ane Pedersdatter; Karen Bæ; Seigne (Pedersdatter) Jensen and Karen Andersdatter

- Because it sure isn't to me -- and tho Geni will let you merge folks who are, in fact, not the same - "what would happen" is you cause a big mess and lots of work for others and/or majorly devalue geni

jeg har nu taget de to familier fra ,

Hans Hansen og Karen Olsen ,

skrevet til oprette @Gunnar Sparre Lorentzen, og forklaret , dog ser hans træ intakt ud

Lars Hansen og Margrethe Hansdatter som tilhører

@Erik Nørring Christiansen

ligeledes skrevet og forklaret ,

mvh marion

Hans Larsen

det som er lagt ind fra denne GEDCOM som er kommet på denne profil , må være grunden til at alle disse fejlfletninger er opstået

jeg har kontaktet @ Mike King , da jeg har forsøgt igen at tage et par fra ,
Lars Svendsen og Grethe Hansdatter disse er fra Bornholm

mh marion

This is a result with gedcom:
Aalov Haraldsdatter

Private User I have posted a separate question here: https://www.geni.com/discussions/201550 but I'd like to ask it here too as this is probably the appropriate thread. I am Dutch.
I am importing a Gedcom file right now. There are a few matches, but the information that the original uploader has entered is mostly less than mine. I suspect that the ancestors of this match also contain less information than what I have gathered over time (as I am Dutch and the other researchers are mainly American so they have more issues to find the information). If I accept the match, then all my information about the ancestors of this match is not downloaded and I have to enter everything manually again. Is my assumption correct? What is the best solution? Not accept and then later merge everything (so at least my work is not "lost") or accept and then manually check all ancestors of the match to see if there is information missing?

Meta Hörz I will follow up on this. I am just returned from a conference in Amsterdam so maybe someone else will contact you.
We have several GEDCOM import experts in the Netherland. I think we are with 3 persons now on the GEDCOM import helpteam now.

On the request of Meta Hörz please use the other thread for comments on her questions.

Showing 91-120 of 191 posts

Create a free account or login to participate in this discussion