Jump to content


Photo

Roots Magic 7 for MAC GEDCOM Generator

GEDCOM Speed

  • Please log in to reply
23 replies to this topic

#21 StephenMWatts

StephenMWatts

    Member

  • Members
  • PipPip
  • 27 posts

Posted 30 March 2016 - 09:53 AM

Renee, I thought I would offer you some more insight on the tree I just sent you. The tree has been built essentially entirely on ancestry.com. My wife has obviously leveraged the work of others just as she allows her work to be leveraged by others, so I can't speak for any of these "outside" contributions. The tree was synchronized with FTM3 for MAC, and then imported directly from the .ftm file. I had also tried the .ftmb file previously and my latest attempt used the .ftm file to see if maybe the importer would make a difference - which it doesn't seem to have. I believe I alluded to the fact that the tree sync portion takes in excess of 12 hours. Ancestry has to "think about it" for a "few" hours (sorry, didn't time it), and then FTM3 has to "think about it" for a "few" hours. I believe FTM3 has to "think" significantly longer. If it wasn't such an ordeal I'd test it out. In the future I could if that would be helpful in any way, but we have to catch it at a time where my wife is willing to give up her tree for an undetermined period of time, because you can't do both.



#22 StephenMWatts

StephenMWatts

    Member

  • Members
  • PipPip
  • 27 posts

Posted 30 March 2016 - 11:24 AM

Zangrou, I was theorizing in my sleep/wakefulness about a metric that might help to determine where problems might occur, and before you messed me up with your "messy" facts, I was thinking record density (read records/person) might be something to be concerned about. Your tree has a density of 10.17 records/person - counting all line items equally (probably not valid, but simply theorizing). The tree I'm working with has a record density of 19.02. Jerry's has a density of 5.16. I would expect trees to vary widely depending upon individual emphasis - which these do. I decided to put these on a spreadsheet for a slightly deeper look - still thinking about whether or not it "says" anything. I have a couple more tweaks, and since I haven't yet figured out how to attach a screenshot or file, I'll tweak it just a bit. Can anyone help me with how to attach a .jpg or .png here?Ah, heck - here's an open link: here.

 

 



#23 zhangrau

zhangrau

    Advanced Member

  • Members
  • PipPipPip
  • 1520 posts

Posted 30 March 2016 - 02:33 PM

To post images on this forum, here's what I do:

1 - Screencap or save the image to my hard drive, and resize to 640x480 or lower

2 - Log in to http://imgur.com and upload my image

3 - Ctrl-C to copy the Direct Link generated

4 - On the RM forum, use the [ ] Image icon (just left of the <> Code icon) to open a dialog, then Ctrl-V the Imgur Direct Link URL

5 - Save the RM post



#24 StephenMWatts

StephenMWatts

    Member

  • Members
  • PipPip
  • 27 posts

Posted 06 April 2016 - 09:32 PM

Thanks for the instructions, Zhangrau. It looked like that was what I was going to have to do but at the time it was too much work. Sounds kind of lazy now...

 

Another user offered to tinker around with my tree. He saw the same thing Renee was seeing with DB errors when first calling up my tree. After using DB Tools, he started tinkering around with settings, and when he turned off the export option "Extra Details (RM Specific)" and then my GEDCOM generated in right around 10 minutes, which is more like folks seem to think it should be.

 

I've been decompressing a bit, and exploring other things but thought I should come back and try to close out the discussion. I don't see that there is a good conclusion here other than that there is something about my tree that triggers a reaction in that option. I haven't analyzed the GEDCOM in detail yet, but it does seems to at least be valid. It has fewer GEDCOM errors than most I've analyzed, and nothing that I can see that would be catastrophic. My main interest - which is a more complex analysis - is to figure out if it has as much more of our data from Ancestry (by way of FTM) in the GEDCOM as we are hoping it will. I'll be pursuing that when I stop exploring RM tools.

 

One GUESS is that it is perhaps trying to perform 404 checks on all of the URLs. With several hundred thousand of those, if it is doing that it would explain the slowdown quite nicely. It makes me wonder how many URLs others have in their trees? If I look at the spreadsheet I generated, It is actually possible. Jerry's reported 1,329 links and 30,540 people/minute. Yours had 75,251 and 4,140/minute. Mine has 695,455 and 5-ish/minute. I don't imagine 404 checks are quick, especially if they fail. So far it simply doesn't rule it out as a cause. If that is involved, it would be doing to 404 checks as it goes along. I'm going to have to do those checks eventually, but I don't see the value of doing them while generating an export... Just a guess but I like it as a candidate. In the mean time I've found another issue so off to another thread.