Multiple, Linked GEDs

PaulDesmondWhite · Post by **PaulDesmondWhite** » 15 Apr 2008 14:58

I have quite a number of GED files for various purposes besides my direct ancestry: mostly one-name studies, but also some sporadics for other interest areas.

It would be rather stupid if all these unconnected data sets were stored in a single GED, and even more stupid (perhaps) if I used one GED for each different surname! Or would it?

Suppose I open White.ged and follow the diagram back to Charles, where you find he married Ada Ambrose. Only her record is a stub... a proxy for a record in Ambrose.ged.

Of course, in my dreams, Family Historian has no problems with that. Quietly it goes off and opens Ambrose.ged, pulling her data into memory and carrying on like that while I navigate up the tree of the Ambrose line. Etc, etc!

Actually, my structures wouldn't need to be as regimented as that. Any old division into a few GEDs would do to start with, but in time certain chunks could be hived off to a separate GED, leaving a proxy in just the right place to make the link.

Ah, if only... I'd never make another wish, promise!

ID:2857

YouTube · Post by **Jane** » 15 Apr 2008 15:08

I am a bit lost as to the benefit of this over using one large gedcom file for anyone who is linked into the main tree in any way.

PaulDesmondWhite · Post by **PaulDesmondWhite** » 15 Apr 2008 15:45

Should i reply or shut up? I don't know the form here, sorry. Some of the GEDs have got many hundreds of names, some will go to thousands. Don't want to send all this stuff up to a web site or to family members of specific branches.

JonAxtell · Post by **JonAxtell** » 15 Apr 2008 16:08

This where is where you can use the Split Tree Helper function to prune the tree as you send it out.

My own gedcom has nearly 30,0000 individuals of which I only publish some 12,000 names. I use the split tree helper function to remove the individuals not directly connected to my research and also to remove extraneous information.

ChrisBowyer · Post by **ChrisBowyer** » 15 Apr 2008 16:14

To reinforce the argument, ours is a similar size, FH has no problems with that. And it means you don't have to maintain the same information in different places, and can spot unexpected connections.

jmurphy · Post by **jmurphy** » 15 Apr 2008 16:15

Jane said:
I am a bit lost as to the benefit of this over using one large gedcom file for anyone who is linked into the main tree in any way.

I tend to use a lot of small files as a sort of 'workbook' to sort things out until I am sure what information actually belongs in the big file. If I want to review the census data on a particular family (that is, one set of parents and their children), I might start a new file and fire up GC and actually re-enter the data from scratch, because doing so forces me to look carefully at the data.

It's the small-living-space equivalent of writing stuff on a chalkboard/markerboard.

For Associated Persons like the godparents and neighbors, I keep in my main file the people who are actually associated with the family. If I have a need to study their families, they get their own files.

And I can see where, if cousins are habitually sharing a file where they are studying their common surname, they might keep files which only have their own spouses as a 'contact point' and not their spouses' lines. It would save one having to use the Split Tree Helper over and over and over and over and over again as they send files back and forth.

In other cases, I might keep information separated by location.

If you have a new, fast computer, sure, I can see there is a simplicity in keeping One Big File. But until one is used to working with thousands of names, sometimes it is easier to work with subsets.

Jan

ChrisBowyer · Post by **ChrisBowyer** » 15 Apr 2008 16:16

And by the way, the form here is, argue as much as you like. Nobody should take offence, and it's always good to hear how other people do things (and why).

YouTube · Post by **Jane** » 15 Apr 2008 16:37

I don't have a problem with lots of files, what I can't see the point of is keeping known linked data in multiple files and wanting FH to link between them on the fly.

PaulDesmondWhite · Post by **PaulDesmondWhite** » 15 Apr 2008 17:03

I know this is a bit cheeky, but that's rather like having a set of normalised tables in a relational database, compared with one great big spreadsheet.

Still, i certainly get the point guys and gals, so thanks for all the reaction. Probably feel a lot more in tune with your approach when i've got more experience. Especially if the splitter gets more features....

ChrisBowyer · Post by **ChrisBowyer** » 15 Apr 2008 17:10

A bit off the point, but I think your analogy is backwards... you can only have normalised data in one database. Having seperate linked Gedcoms means you have the same information (at minimum, the name of the linking individual) in two places instead of one.

jmurphy · Post by **jmurphy** » 15 Apr 2008 17:53

Jane said:
I don't have a problem with lots of files, what I can't see the point of is keeping known linked data in multiple files and wanting FH to link between them on the fly.

I agree with you on this point -- if you want them to be linked together all the time, that's a good sign it is time to merge the two smaller files. [wink]

Jan

Post by **NickWalker** » 15 Apr 2008 18:31

As Chris said it is far easier to keep the data normalised if all the data is in one file. Having said that most database management systems do allow different databases to be linked. I suppose one reason you might split into different files would be if you had multiple users on a network with different access rights because you want to keep some of the data private from some users, e.g. you had one file that all users could access, another that just some could access, another which only you were able to write to but was read-only to others, etc. Other than that scenario I agree with Jane, Jon, Chris and Jan.

Family Historian User Group