upload image

Splitting GEDCOMs - The Targeted Approach

Privacy Level: Open (White)
Date: [unknown] [unknown]
Location: [unknown]
This page has been accessed 154 times.

Contents

Splitting GEDCOMs - The Targeted Approach

When thinking about approaches to splitting GEDCOMs

Strategy

Our strategy is to follow a single line of ancestry, following family husband-wife pairs based on the male line of descent. Consider this table of the line to my grandfather (Elwin, but he went by "Bud") from my 4th great grandfather, founder of my Ebaugh family line, Joseph Ebaugh.

Ebaugh Line

The filters that I am teaching you will capture:

  • Children of each pair.
  • Spouses and children of these children.
  • Siblings of the husband and wife in each pair.
  • Children and spouses of each sibling.
  • The spouses and children of the children.
  • The parents of the wife in each pair.
  • Some multiple spouses.

They will not capture:

  • Parents for any spouses other than the wives in the main line.

Occasionally the filters miss spouses for unknown reasons. For example, it always leaves out my grandma Ebaugh's (was Harbeson) sibling's spouses, although it captures their children! It is trivial to create a GEDCOM with just the missing spouses, but it's still a little irritating.

To choose our endpoints, we start from someone we've already added to WikiTree (or the home person on your tree if you haven't entered anyone yet). We choose our end point by following our tree back until we determine where in the line we're examining we encounter an individual ALREADY on WikiTree. Most of my lines that did not terminate in 19th century immigrants hooked into the "world tree" around the 4th or 5th grandparent level. We do a little more work now identifying where this occurs to save a lot of frustration later on, when we have to go through the GEDCOMpare process. I'll go over an example of choosing an end point later on, but for now the Ebaugh line terminates in a 19th century immigrant whose parents we have no knowledge of, so Joe gets the endpoint trophy.

The Genesis Filters

Some of you probably skipped right to this part. You might want to look over some of the other stuff I've written here. There's even jokes! Or maybe it's because of the jokes that you skipped ahead. Or maybe the screen shots. Well, after this point there are no more screen shots. I assume you've grabbed the pebble from my hand, grasshopper, or in GRAMPS filter building terms, you have a basic idea of what's going on.

These filters are used the first time you're adding people to the tree. In the beginning. (Get the name now?) it assumes your start person's and descendants' families haven't been added to WikiTree yet. That may be what you want and you expect duplication, just beware that that will happen.

Filter A: The Line

This filter creates the line of ancestry between the two families you've chosen as endpoints.

Ancestral filters-->Ancestors of <person>, choose your start point (the person you are moving backwards in time from); Inclusive=yes (click the box).
Descendant filters-->Descendants of <person>, choose your end point (person you are moving back in time towards); Inclusive=yes.
All rules must apply

Filter B: The Spouses

This filter brings in the spouses of the males in the line. It is necessary to capture multiple spouses for males in the line. I haven't yet figured out a way to capture them automatically for others. At this point that is still a manual task.

Family filters-->Spouses of <filter> match; filter A
General filters-->People matching the <filter>; filter A
At least one rule must apply

Filter C: The Siblings

This filter brings in the siblings of both spouses.

Family filters-->Siblings of <filter> match; filter B
General filters-->People matching the <filter>; filter B
At least one rule must apply

Filter D: The Parents of the Wives

This filter brings in the parents of the wives and their siblings. It is necessary because siblings without parents must be handled in a special way on WikiTree. If the parents are available, we should include them. If the parents are legitimately not available, the siblings can be handled by including the relationship in their bios.

Family filters-->Parents of <filter> match; filter C
General filters-->People matching the <filter>; filter C
At least one rule must apply

Filter E: The Descendant Family Members

This filter brings in family members, spouses, children, the family members of the descendants of all the people we've explicitly added thus far. It's impressive. Running filter D on my tree yields 74 people, filter E gives up 273! It does miss the occasional spouse here and there, but it's nothing that can't be quickly taken care of. Notice that we feed the descendant family members rule filter C, NOT filter D. We do, however, pull from filter D to get the parents.

Descendant filters-->Descendant family members of <filter> match; filter C
General filters-->People matching the <filter>; filter D
At least one rule must apply

Searching WikiTree to Find Endpoints

I'll be using an example here. When I ran the filters to make the first GEDCOM, my Great-grandmother Nellie Pitts was one of the endpoint people. The Pitts line is long in my family tree, so I'm going to follow the Pitts back in time starting with Nellie.

On WikiTree hover over 'Find' at the top of the screen and choose 'Search' from the dropdown menu. On the search screen enter the surname you are interested in following, in this example 'Pitts'. On the 'PITTS Genealogy' page that follows choose 'date order'. This will take you to the 'Pitts Genealogy' page. On the right had side in the box titled 'Change this list' choose 'Sort with most-recent birth dates on top'.

Look at your family tree (the one you're trying to put on WikiTree) to find the name and birth date of your source person's father. Scroll down through the list in WikiTree to see if he's there. If not, scroll through looking for his father. Repeat as necessary.

When I went looking for fathers I went through George Prewitt Pitts (1869), John Luther Pitts (1825), and John Henry Pitts (1798) without finding anybody already on WikiTree. But the next Pitt, John J., was on WikiTree! We've found our end point person - John Henry! Not John J, if we picked him it would end up duplicating people on the tree. If there are people to be added around John J. other than John Luther,, better to deal with them separately.

You're probably wondering why I don't just search for specific names. I find it easier to just lean on my down arrow key rather than typing words. You mileage may vary.

Filters After the Feasting from the WikiTree of Knowledge

Now that we've found our endpoint person and have that knowledge, we can construct filters to isolate people we need from our tree. These filters are very much like the set we've already developed. So much so that I'll just summarize them here for convenience, and point out the difference when we encounter it.

Filter A:

Ancestral filters-->Ancestors of <person>, choose your start point (the person you are moving backwards in time from); Inclusive=yes (click the box).
Descendant filters-->Descendants of <person>, choose your end point (person you are moving back in time towards); Inclusive=yes.
All rules must apply

Filter B:

Family filters-->Spouses of <filter> match; filter A
General filters-->People matching the <filter>; filter A
At least one rule must apply

Filter C:

Family filters-->Siblings of <filter> match; filter B
General filters-->People matching the <filter>; filter B
At least one rule must apply

Filter D:

Family filters-->Parents of <filter> match; filter C
General filters-->People matching the <filter>; filter C
At least one rule must apply

Filter E:

We make one small change in this filter to deal with what would be a problem. The rule 'Descendant family members of <C>' would match the descendant family members of our start person, which we already matched in our first set of filters. The solution is simple: add a rule to match the descendant family members of our start person, and set the filter to 'Exactly one rule must apply'. This prevents the start person from matching, since they (and their descendant families) will match TWO rules.

Descendant filters-->Descendant family members of <filter> match; filter C
General filters-->People matching the <filter>; filter D
Descendant filters-->Descendant family members of <person>; chose your start person, inclusive

Exactly one rule must apply

Now go forth and filter! Or read the next part (when I finish it).

Splitting GEDCOMs With GRAMPS
Importing a GEDCOM into GRAMPS
Using filters in GRAMPS
Splitting GEDCOMs - The Targeted Approach



This is an "orphaned" profile — there's no Profile Manager to watch over it. Please adopt this profile.


Collaboration
  • Login to edit this profile and add images.
  • Public Comments: Login to post. (Best for messages specifically directed to those editing this profile. Limit 20 per day.)
Comments: 1

Leave a message for others who see this profile.
There are no comments yet.
Login to post a comment.
Filters ABCDE are a great start, thank you. Having a huge tree to import, I have developed a couple more which narrow down the results and cut down the number of duplicates to process in GEDcompare:

Filter F - Limits the results of Filter E to a single family surname: People matching the <filter>. Filter name:="E" People with the <name>. Full Family name:= "[choose the surname you want to work with]" All rules must apply

Filter G - Limits the time range when the people you are working with were probably alive: People matching the <filter>. Filter name:="F" People probably alive. On date:="[pick a year]" All rules must apply

posted by Edward Young