Best practice on multiple Gedcom's
Brought to you by:
canajun2eh,
yalnifj
Menu
▾
▴
So, I'm relatively new to PGV. I can see that the software supports multiple Gedcom's beneath the surface. I can hypothesise about good reasons to do this but I'm interested in hearing the voice of experience. Was it worth doing?
It will depend on how you want your site to work. There are actually two features involved - multiple gedcoms; and linking gedcoms (either on same site or remotely)
For me, the answer is certainly Yes - very much worth doing.
My tree was getting large and unwieldy - with a couple of major branches linked to my own family by only one marriage. So, I've split into three trees, and used linking to maintain the look of connection. This has left me to focus on my own tree, while other's focus on theirs. They can control access, default theme etc etc.
As well, I have a 4th tree not connected (yet) to the others. It has people in it that I hope I will one day find a connection to.
I now have a policy - if a new tree is greater than (say) 20% of my own, then I will add it as a new tree. If less, then I will go to the trouble of merging it.
Merging was another key factor for me - its hard to do properly (whatever method / software you use)- and linking separate gedcoms makes it unncecessary.
Two questions are raised by this:
1) Should each gedcom have its own sql database?
2) Once databases are linked, should new children to linked individuals be entered in each database?
I have about 30 GEDCOMs to load and trying to get the configuration and privacy the same on them all is a pain.
Is there any way of ensuring Configuration and Privacy on all GEDCOMs is the same?
I have looked through WIKI, documentation etc and this is ther only place I can find that gets close to my problem.
I am running a one name society website www.buntingsociety.org.uk just in case you were wondering why I should want to load so many GEDCOMs.
Jerry Green
I love the idea of splitting up a GEDCOM into multiple files. You say that you have separate access, themes, etc... but can you still do the same reports and searches as if they were in the same file?
Also, I assume that you gain a significant performance improvement if your file is divided into two halves rather than as one large whole one (especially once you get over 20K names in each file). From reading other threads, I know that PGV handles large files, but I'm curious if anyone has figured out at what point it would be better to split it into multiple files. (When it gets too big, I have been getting out of memory errors)
Hi,
let me explain my experience.
One day I recieved four Gedcoms to be connected to my own one. I started merging them with comparing each overlapping record. I used PAF for this comparison, I wrote the recordnumbers down and then merged them manually because I was very suspicious let PAF do it automatically.
Now the foreign Gedcoms are not updated only my own Gedcom but they are merged. If I receive one day an updated version of some of the others I have to merge them again. I don not want to have this work again.
Now since the feature of linking Gedcoms in PGV is available I am thinking about of parting the Gedcoms down to their original structure and link them exact as nigelos approach. I think this is a real valid reason of using multiple Gedcoms and keep them in sync, see the discussion of "keeping offline Gedcom in-sync" as well.
Hi,
I am also new to PGV and love what I see so far. However, I am unclear as to how to best split up my GEDCOM.
My original GEDCOM included my wife's tree and her side of the family took over my 'welcome page' statistics because we have more information on that side. So I wanted to split my original GEDCOM so that each side of our family can login and see statistics related to their side of the family.
I used my offline family tree app to split my GEDCOM into two GEDCOMs. One with me as the root and my side of the family. I then did the same making my wife the root with her side of the family. The third GEDCOM was me, my wife, and my two kids.
I then uploaded my file, uploaded my wife's file and linked her tree to mine as my wife. This seemed to work fine although I wasn't able to see her parents or any of her family tree information when looking at my GEDCOM. Should this be possible?
I then linked my third GEDCOM (with just my immediate family) by linking me to the me in the third GEDCOM. This created a second family in the "Close Relatives" page, but had no information for the family.
So with these little problems, I'm not sure if the linking is working incorrectly or if I'm not using it as it was intended. Could someone please tell me the how the linking was designed to work?
If you've gotten linking to work well, can you please tell me how you do it?
This is a great app and I look forward to announcing it to my family to have them fill in my missing information. Keep up the good work!
Norton
While I haven't tried it yet, the ability to link gedcoms was one of the reasons that I got excited about PGV in the first place. I hope that this capability will help me to better accomodate the extended family.
I just searched the forum archives for "linking gedcoms" and found numerous useful posts that have addressed the issue in the past. These were quite helpful. I wonder if someone could name a working site with linked gedcoms and point out the details of one of the links so that we can see what it is possible.
With upgrades to the wiki, the links mentioned in several of the other threads on this topic are broken. Here is the current link to "How To:Remote Link Individuals Across Websites And Databases"
http://wiki.phpgedview.net/en/index.php/How_To:Remote_Link_Individuals_Across_Websites_And_Databases
(That is the hazzard with linking to a wiki that has a changing structure)
Linking GedComs, yes this is what I wanted to do. Unfortunately I can't find documentation on how to do this. Where is the documentation/help information? It is not embedded in the software. Why?
Norton --
I haven't gotten this to work either. It was my impression that once you create links of individuals in separate files, it appears almost as if it were one big file again. (How this works, it isn't clear)
But my most pressing problem is because my two gedcom files are from separate families with overlap of over 20 individuals (I'm guessing). Rather than linking every individual that appears in both gedcom files, and maintaining those links with each upgrade, I'm guessing that it would better to remove the duplicate individuals from one file just so that there is only one or two records in common between the files.
-David
After trying this out, here are a few things that I have found.
1. Most all of the reports/functions treat linked gedcoms as if they were one big file. The ones that don't are the gedcom specific stats, lists, and calendar functions... but that seems to make sense why these would not be included.
2. The links are stored in the gedcom file. Because of this, they work one way, from an individual in one gedcom pointing to a record id in another file. If you want it to link back from the second file to the first one again, you need to create a second link. (If they are both local gedcoms, I don't see a reason why you wouldn't want to link both ways)
3. If you need to upload a newer version of a gedcom to your site, it will overwrite any links which you created in that file. (I'm guessing that links in other gedcoms which point to this one will still exist, but I haven't tested it yet.)
4. Using offline software like PAF Insight (http://ohanasoftware.com/) you can more accurately see which records are similar between the two files, and then create the links in each respective online gedcom.
5. There is some logic which tries to determine if the family of the person being linked exists in both gedcoms and automatically links them as well. I haven't been able to confirm this to work well, because in my test it only linked the one individual. This is probably due to the fact that there was enough differences in the data between the family members.
6. Creating separate gedcoms means that different users can be assigned to administer each file differently. General users can be granted different rights to each respective gedcom file based on any number of criteria.
7. If you find that there are a large number of individuals which are in common between gedcom files, you will have to weigh the cost versus benefits more carefully before creating a large number of links.
Considerations: total number of links, one way versus two ways links, frequency of gedcom updates (overwriting links), need to administer separate rights of users to view/edit data in each gedcom, total number of records, ... others?
On my site, I have thousands of records in the main gedcom and a little over a thousand in another one which I received from other family members. Digging into the data, I found that there are several hundred individuals which overlap in the two files, making creating and maintaining the links difficult. Considering that significant work is still being done offline and new gedcom files will need to be merged with online changes and uploaded on occasion, it would mean that the links would have to be recreated regularly. The approach of just merging the two files into one, rather than to link the two files, has merit but I'm still not certain. If I can convince family members to use PGV exclusively, it would be worth the effort to create the links.
Reading all these entries is very interesting. I am new to GEDCOM and phpGedView, but I like what I see! After inputting my first 30 (yes I am only at 30) family members last night, I realized that this could become quite the massive file. If left as one gedcom file there would be a lot of data from one branch of my family that the other branch of my family could care less about. So now I am considering splitting the data 5 files between my wife and my family.
At the moment my thoughts are to take the major Surnames and turn them into individual files. If there are other surnames that come to a dead end or that we don't know much about at this time, I am planning on grouping the together with the major surnames.
My question is two part:
1) Does this sounds like a good way to organize or would you recommend actually creating individual files for every surname in my family?
2) If in the future we find more information on a different surname, how easy would it be to split that surname to another file? It seems like it would be confusing. Is there documentation on this?
I have a question related to this, too. I have two GEDCOM files. One for me and my family and one for my wife and her family. We got married now and so I linked my wife from her GEDCOM to me in my GEDCOM as my wife over this link to external person feature. Is it right that then there will be created a new individual in my gedcom for my wife? So far it seems to work and then I added also me as her husband in her GEDCOM, but now I got confused about the familie that we created now. Cos there exists also now two families for us. One in my GEDCOM and one in her GEDCOM file. Can I link the families also somehow or must I enter these data twice? But there seems also to be some other problems cos when I go to the new automatical created individual in my GEDCOM file for my wife she is shown to be married twice. Do I need to link the people in different way like not about husband/wife or any other tips how I could fix it. Thank you very much.
Shane, my first thought is that you should not yet worry too much about size - you have a long way to go. However it is a good idea to plan ahead.
My main gedcom is around 5,000 people, and that causes no problems at all. The fact there are branches not everyone is interested in is largely irrelevant - they just don't look at them! I started using multiple gedcoms and linking when my main file got to about 6,000 people - but mainly so I could keep myself focussed on a 'smaller' group.
A file for every surname would make your system extremely complex, as each gedcom has its own configuration and privacy settings, each user has different access's, etc etc.
My advice would be keep it simple (one file) for now. Splitting later is not hard. I use a piece of free offline software, gedsplit.exe. My own preference is also to only split where there is only one link, or at least as few as possible. A marriage is the most obvious.
My 2nd suggestion, while you are sensibly thinking ahead, is set yourself some basic rules, like:
1 - be consistent and complete with place names. Place Hierarchy, maps etc are useful tools, but only if you are consistent with formatting places, and thats easier when you start than when you have entered a thousand records.
2 - enter ALL your data as you go. I'm now regretting not entering census data fully as I researched it - I now have hundreds of pages to go back and enter!!
3 - think about your media storege. Once you have a couple of hundred media items, it gets hard to move them into a more organised file structure than a single folder.
I wish I'd followed my own advice - so here's your chance to avoid my mistakes. Good luck.
Also as someone starting out, make sure to source every bit of data thta you have. This is a regret that I have when I come accross a fact in my GEDCOM that I forgot to source long ago.
Thank you for these great tips. Question about the media files - little off topic, but I wanted to see how you have things set up. I read the help file and it suggests keeping the media somewhere not Internet accessible, so I placed the media here ../../genealogy/media/ and I changed the media directory levels to 1.
The photos are being added to the directory media and media/thumbs directory, but they are not being displayed on the Internet. Additionally the help files explains that with levels set to 1, surname directories would automatically be created if they didn't exist during upload. Instead it created a media/media and media/thumbs/media directory. Have you ran into th is problem? Are you automaticaly creating surname directories?
That note in the help file is wrong. It has been noted and will probably be corrected some time.
You can't store media outside of the pgv directory. In fact it needs to go in the pvg/media folder, or a sub-folder of that.
(Having said that, it is technically possibly, using 'symlinks' but thats another story - search these forums for info on that).
Medai folders are not created automatically, you need to set them up carefully before you start loading images.
I use a single level structure with different folders for each gedcom and for photos and certificates, so the structure looks like this:
pgv/media/gedcomA
pgv/media/gedcomA-certs
pgv/media/gedcomB
pgv/media/gedcomB-certs
You could add another layer, like this, if you want a more detailed structure:
pgv/media/gedcomA/photos
pgv/media/gedcomA/certs
pgv/media/gedcomB/photos
pgv/media/gedcomB/certs
The appropriate matching thumbs directories are then set automatically.
All this is done in gedcomconfig, and / or media management. You need to set media directory levels to 1 or 2 (instead of default 0) in the gedcom config.
The best explanations for all this are in the wiki pages.
"That note in the help file is wrong. It has been noted and will probably be corrected some time.
You can't store media outside of the pgv directory."
I do. Well, I guess I don't know what you mean by PGV directory.
My index file is outside of the directory containing config.php and my media is stored in a subdirectory of the index.
I have linked two gedcoms on the same site.
When I ask for a Relationship Chart between a person on one chart and a peron on the other ...NOTHING.
I can input the ID of the person whose gedcom I am in but cannot even see the other person.
Given that this Relationship Chart was one of the main reasons I decided on PGV I severely browned off.
Question1 - No, to have separate databases for each gedcom would also require two complete installations of PGV (in particular the config file which sets the connection between pgv and the database), which defeats one of the key reasons for same-site linking. The table structure of the pgv database recognises the existence of separate gedcoms.
Question 2 - also no. Adding the children to the 'master' gedcom (the one you link TO, rather than FROM, should show links to all related people.
Jerry
I havn't trtied this myself, but it should work OK. Just be careful to store backups as you go.
Set up one GEDCOM as you want it to be.
Import the second, but don't attempt to alter any of the GEDCOM or PRIVACY settings.
Once it's imported, go the the /index folder. Each GEDCOM will have two unique files (plus others, but these are the two that matter):
GEDCOMXXX.ged_conf.php and
GEDCOMXXX.ged_priv.php
These are where all those settings are stored. Simply copy the two from your first GEDCOM, and rename them to match the second GEDCOM. You can then repeat that as often as you like.
Thanks for the hint, that seems to work great.
I notice that these two files in their headers say they are default configuration and privacy files. I looked around and they seem to be based on config_gedcom.php and privacy.php.
Can I edit these two files to my default values so that when GEDCOM is imported no set up is required?
Quote"Thanks for the hint, that seems to work great.
I notice that these two files in their headers say they are default configuration and privacy files. I looked around and they seem to be based on config_gedcom.php and privacy.php.
Can I edit these two files to my default values so that when GEDCOM is imported no set up is required? "
I have now loaded five gedcoms using above and it works great, however it is rather annoying to have to set up new Welcome pages, Welcome nlock and Portals for each GEDCOM. Is there any way of keeping all of these consistent with each other for all GEDCOMS? I still have at least another 25 GEDCOMS. The Website is for a one-name society.
Yes, I don't see why not. Probably a better idea than my suggestion. Anyway, it seems you have plenty of GEDCOMs to experiment with. Let us know how it goes.