US and DE date formats mixed up in Ancestry GEDCOM import

5.1.5, Windows 10

I have imported my family tree via GEDCOM from Ancestry, and now see some dates printed in US format yyyy-mm-dd, and some in german notation dd.mm.yyyy
I tried to change this manually in GRAMPS, but the change is not accepted.

Any idea ?

Walter Fertl

2 Likes

Hallo Walter,

How do you mean that the change was not accepted? Did you try to edit the wrong dates? Or are there too many to do that? Or did you try to change the date format in the Gramps settings, and didn’t that give you what you hoped for?

I’m in The Netherlands, and I sometimes see dates coming from FamilySearch that were typed in a way that FamilySearch did not understand, but the number of such dates is often small enough to be corrected.

Can you show a piece of your GEDCOM file here, and maybe also a screenshot of your date settings? I can read German, and it helps to see what’s really going on.

Viele Grüße aus Driebergen,

Enno

Hi Enno,

what a fast response, many thanks !

Well, there are many persons and events listed in US format. I tried to change a few persons, but when looking to the person’s record after the attempted change, it was the same as before.

Also, I saw that within one and the same person, the date notations are sometimes different, and some persons are printed in normal, some in bold.

You mentioned there is date setting option in Gramps. I am a rather new user, can you point me to that ? Maybe that solves it already ?

If that does not work, I am happy to send you part of my GEDCOM download and screenshots.

Viele Grüße aus München,

Walter

1 Like

Welcome

Sometimes the dates on a web site are stored as text entries and not pure dates. When they are imported into Gramps, Gramps will accept the text entry and store them in the event. These text dates will often display as bold or in a hodgepodge of styles.

If you were to create a new event in Gramps for a person, you enter the information as a text entry and in the save process, Gramps will convert it into the date entry.

Did you edit and save the person or did you edit the Event record attached to the Person’s record. Every component within Gramps is its own record which allows for sharing.

Hallo Walter,

When you install Gramps, the default setting for dates is the ISO format, yyyy-mm-dd. You can change that in the 3rd tab of the preferences dialog, which looks like this in Dutch:

I made this screenshot with the date notation expanded, so that you can see the choices that are available. The dialog will have different words in German, but the position of items is the same in all languages, so I bet that you will understand this one anyway.

When you change this, and restart Gramps, most dates should appear in the format that you chose, which means that Gramps knows that they are dates, so that it can display them in the right way. It will then display Mai where I see Mei, and English users see May, assuming that you chose to display (abbreviated) month names, not numbers.

In your case, you will still see dates in bold, in the format that they have in the GEDCOM file. Gramps does this for every date that it didn’t recognize as a date in the GEDCOM file, meaning any date that doesn’t follow the standard format, which is 1 MAY 2023 for today.

If there are not too many dates that are wong, it is easy to retype them in the proper format, which is the same format as you chose for display. If there are many, it might be easier to load your GEDOCM file in a text editor, like notepad, or notepad++, and do a global replace in which you might need to replace März with MAR. You may need to be careful with that, if you also have these month names in notes, for which I assume that you want to keep those in German. GEDCOM month names are 3 letter abbreviations in English, so if your GEDCOM has Okt(ober), you must change that to OCT. Same for Dez(ember), which must be DEC in GEDCOM.

If you’re unsure about the number of non-standard dates, there is a quick way to find them, and that is by choosing the Events category, and sort that by date. This will show all empty dates first, followed by all dates that are wrong, displayed in bold, and the good ones last.

This is for simple dates, and you may also see problems with dates before and after (BEF and AFT in GEDCOM), and date ranges. And again, if there are just a few, you can simply retype them, with the proper keyword(s), in German, which might be bevor for Before, and nach for after. That’s just a guess, because I run Gramps in Dutch, and don’t have the German version installed.

I hope that this helps to get you started, and if not, you know how to find me. :slight_smile:

groeten uit Driebergen,

Enno

Hi Enno,

your first suggestion was rather successful, I now see all dates in german notation, some with the month spelled out, some as a number, however.

And in some cases the month and day are interchanged.

But this already looks much better. Have not found yet the setting for the month format.

The BEF and AFT interpretation seems to work well.

Maybe I should re-install Gramps with -en language and US date format setting before importing the GEDCOM data ?

Thanks for your help, will let you know how I get on.

Regards

Walter

Hi Walter,

If you see inconsistencies, like some months spelled out, and others as numbers, it means that some dates are still not recognized as dates, and are stored as text. And you should see those when you sort all events by date, in the event category view.

And if month and day are interchanged, there may be dates that were stored in US notation, like 5/1/2023 for today. Can you compare those to what you see on-line, and how they appear in the GEDCOM file?

There is no need to add English by re-installing. Most of Gramps is written in US English, so the language support is always there, and GEDCOM import is designed to be independant.

Regards,

Enno

A minor oddity in the Windows AIO installer is that it defaults to UK English in the installer. This despite the development specifying the US dialect.

I do not know if it could cause any weirdness. But rather than take that chance, I always add a US english selection during install.

1 Like

Do a simple test.

Go to the Event category view. Select an event that has date information that does not display according to the date format you selected in Preferences.

Double-click the record to edit the event, then simply save the record. The date information should now be displaying correctly.

If this simple save of the record did not correct the date, let us know and we will try something else.

If the simple save of the record does indeed correct the date you can do the same for all the other events with the irregular dates.

A hint. You can select a batch of records, right-click, select edit. All of the selected records will open their edit windows. Then you can click trough the list saving each record. Doing this batch edit only works easily (for the user) if you are doing the same edit to each record. I would not exceed doing more than 20-25 records at a time. Doing more than this and Gramps will have a memory moment.

1 Like

Hello Dave,

I assume that the GEDCOM file produced by Ancestry, which I imported, contains proper date formats. I checked a few of the questionable dates in Ancestry app (the GEDCOM source) , and these looked ok there.

Trying to change a date, I clicked PERSONS, then doubleclicked the name in the list, then EVENTS and then the questionable date. Gramps allowed me to change the date, but after retuning to the persons’ listing, the date was unchanged.

What did I do wrong ?

Thanks for your help

Walter

Hi Dave,

I tried your suggestion, unfortunately the problem remains.

Regards,

Walter

Hi Enno

I just deleted and re-imported the GEDCOM, and saw that the import reported some 18,000 errors.

Looking up one example of mixed up date formats in the GEDCOM file:

1 NAME Werner /Fertl/

2 GIVN Werner

2 SURN Fertl

1 SEX M

1 FAMC @F85@

1 FAMS @F22@

1 BIRT

2 DATE 10.9.1941

1 DEAT

2 DATE 14.11.2016

Birthdate and dead date seem to be in the correct format.

However, at Gramps these look different:

… continued in next Email

They are are NOT in the correct format for GEDCOM. Valid in GEDCOM uses the dd MMM yyyy format (or the title case dd Mmm yyyy in GEDCOM 5.4 variant) and specifies English. As before, the dd.mm.yyyy is considered ambiguous unless the dd value is greater than than the 12th day of the month.

This is a continuation of the invalid imported date re-parsing conversation in the following thread. (Although they were dealing with ambiguous dates from an Italian rather than German GEDCOM.)

And this issue has the same complication that this will have a short-lived interest in resolving the import problem. (Once this particular import case is resolved, interest in evolving the tool from a crude hack to an elegant and flexible tool will die. So we have to take advantage of the short-lived interest to evolve the old solution hack, not start all over again.)

Perhaps this cycle of the evolution could pick up the development thread? Maybe add the ability to list examples of invalid dates from the database and offer the user and option to specify which decimal delimited pattern (‘dd.mm.yyyy’, ‘mm.dd.yyyy’, ‘yyyy.mm.dd’, ‘yyyy.dd.mm’, or the no-leading-zero ‘d.m.yyyy’ ‘m.d.yyyy’ or maybe even the evil ‘d.m.yy’ with a century selector) to use for the re-parser?

Again, using the PortableApps fork on a Windows box (or emulator) was the starting point because it includes all the languages Gramps supports.

The two dates in the GEDCOM appear to be using a d.m.yyyy format. This based upon there being no 14th month.

But even the birth date that Gramps is displaying in a valid format appears to be an error.

To be clear, is the birth date 9 October 1941 or 10 September 1941?

I fear you will need to do manual edits. Both for formatting and accuracy of the date from Ancestry.

This unless there is some way of forcing Ancestry to use valid dates in the export.

Maybe there is a way to re-interpret the date with Python’s locale module? Maybe with an override to force the locale from GEDCOM’s LANGUAGE_ID

The date format in my Ancestry GEDCOM download does indeed not conform to the latest 5.1.1 standard. I have asked Ancestry support for a statement.

However: the birth and death date formats in the supplied example are exactly the same, so why does Gramps show these in different formats ?

P.S. the birth date of the person in the example is Sep 10, 1941

P.S. I have done some Python programming many years ago, I am afraid trying do work within such a complex application may be too much for me.

Hi Walter,

We have two problems here, and the 1st is that your dates are all wrong. They look right to you, but they’re not right in the GEDCOM sense, and this suggests that Ancestry never understood what you typed, because if it did, it would have exported these dates as 10 SEP 1941 and 14 NOV 2016.

The 2nd problem is, that Gramps should not try to make guesses for dates like 10.9.1941, because they mean different things in the US and in Germany, which in this case means that it’s corrupting your tree. And that is very bad for a program with an international audience.

In this case, the solution is quite simple though, as long as you haven’t made any changes in Gramps, because if you’re consistent, you know that all dates have their month number in the middle, between two decimal points. And if you don’t want to program, you can simply edit the GEDCOM file and replace all occurances of “.1.” with " JAN ", “.2.” with " FEB " and so forth. This will work quite well if there are no numbers surrounded by decimal points in other places, like in notes, and if you take care to use the right month names, which are

JAN | FEB | MAR | APR | MAY | JUN | JUL | AUG | SEP | OCT | NOV | DEC

as I just copied them from the GEDCOM standard.

You can of course also write a small Python program that reads the GEDCOM file, copies all lines to another file, and only does this search and replace for lines that start with n DATE, where n can be any number, most probably always a single digit one.

I just asked ChatGPT to write such a program for me, so if you give me your email, in a private message, you get a Python program in return.

Enno

Quite logically, the opposite way does not work, which I experienced first, and then wrongly interpreted that Gramps does not accept a date change this way.:smiley: :heart: :heart: :heart:

The bottom line.

Ancestry is not exporting dates using the GEDCOM standard format. A format that importers agree upon and understand.

Could it be that Ancestry does not user taking their databases off their platform.

Hi,

just got this reply from Ancestry support: the most actual GEDCOM standard is 7.0.11, which is the only standard they support.

Any idea how to import from that version ?

Regards

Walter