Automatic uploading of References to Lost Cousins

So to reiterate, Ancestry gedcom has a different format to that produced by FTM. Alexander has offered to look and see if that format of the "non breaking space" can be added to FTA.

So yes, Ancestry adds the source citation and that is passed successfully to FTM and the FTM gedcom works correctly in FTA. While Alexander is testing you could just use the gedcom from FTM?

As a side question, why do you sometimes use FTM gedcoms and other times Ancestry gedcoms?

 

Did you sort out your male/female issues?

 

That was it I wasn't stripping out the &nbsp field. Fixed for v7.7.0.

 

I also only use FTM generated gedcoms. And yes they do take a long time to load.
Is there a significant difference in file size between the 2 gedcoms?

 

Odd I'd have expected the opposite. Ancestry screws up notes and writes out the data in an inconsistent format that doesn't adhere to the GEDCOM standard. So there is an option to retry notes looking for bad line breaks (the line breaks are the issue). Whereas FTM writes good clean GEDCOM files and should process much more smoothly.

However it may be that the option to retry lines for Ancestry is still on and it's retrying for FTM files when it doesn't need to and that's slowing it down. At present that's just a guess. I can try with my own file later. I'll sync to Ancestry and download an Ancestry GEDCOM and compare that to an FTM GEDCOM. I can run the two under a thing called a profiler that tells me how much time it's spending in each part of the code. This then shows up which block of code is taking the most time for each file. This may provide the answer although as I say it might be down to option settings so I might not get the same results. I'll get back to you if I have questions on what options you have set.

 

Ok I've done initial investigations and can confirm Bob's findings. It does run faster with an Ancestry GEDCOM vs a FTM GEDCOM. This appears (early days so not confirmed) to be down to amount of time spent searching for census references. Incidentally if you turn off the option to search for census references both load pretty much same time.

For some reason yet to be fathomed the Ancestry version seems to create less records that need searched.