[Lifelines-dev] Extracting a usable proper subset of a database?

[David W. <da...@ca...>]

I'm a bit new here, so redirection of my query will be accepted cheerfully.

I've been using LL since mid-April (which corresponds fairly well
with my awareness of the need of some measure of discipline in
researching genealogy-related history), and generally find it very
useful; the scripting capability alone is amazing.

For lack of an obviously-superior approach, I have been using a single
database for the information for both my side of the family and my
spouse's.  I received a considerable "head start" because someone I
didn't know had been researching one of these branches (though the
closest he had come to my immediate family was my gg-grandfather).

Since then, I've uncovered quite a bit of material in various places,
and have relayed the parts I think of to him... but the approach is
awkward and prone to omission.

So I thought I would merely extract a subset of my database -- just
getting the part that is relevant to the one branch -- dump it to GEDCOM
format, and send that to him.

The "partition.ll" script comes close, but is a bit too aggressive for
this purpose:  it puts eveyone who has any documentable relationship to
a given person in the same partition as the person in question.

The description of the "fdesc.ll" script appeared a bit closer, but I
couldn't figure out why, after asking for a person, it then was asking
about a "list of people" (from which only a single person could be
selected).  So I started hacking at the script....

I got it to the point where it would prompt for a person and generate
the GEDCOM file, but trying to read in the resulting GEDCOM caused
errors to be generated:  it seems that some of the siblings of included
spouses were referenced, but no INDI record was present for them.  So I
hacked a bit more, and took care of that.

But now I still get errors (that were in there originally; I just hadn't
got around to mentioning them 'til now):  in the case of records that
contain cross-references to SOUR records (such as @S1@), the references
are in the records in the generated GEDCOM file, but the SOUR records to
which they refer are not.

As a result, the generated GEDCOM (still) cannot actually be used by LL.
(Whether or not it can be used by something else, I don't know.)

I could post-process the GEDCOM with "grep -v," but that seems ugly, and
it removes (potentially useful) information.  I saw mention of the
reference() and dereference() functions, but it's not clear to me if
they might be useful, or if so, how to make use of them.

Baasically, I don't see how to go about making the subset-GEDCOM
sufficiently consistent that LL can actually use it.

Would someone care to provide a hint, pointer, or clue?

I'll append a uni-diff showing my hacks (to date) to fdesc.ll after my
.sig.

Thanks,
david
-- 
David H. Wolfskill				da...@ca...
Trying to support or use Microsoft products makes about as much sense
as painting the outside of a house with watercolors.


--- /usr/local/src/lifelines-3.0.17/reports/fdesc.ll	Tue Nov 28 13:39:45 2000
+++ fdesc.ll	Tue Jul  2 14:11:19 2002
@@ -23,10 +23,15 @@
                 addtoset(set2, indi, n)
                 set(set1, descendantset(set2))
                 set(set1, union(set1, set2))
+/*
                 getindiset(set2)
                 set(set1, intersect(set1, set2))
+*/
                 set(set2, spouseset(set1))
                 set(set1, union(set1, set2))
+                set(set2, siblingset(set1))
+                set(set1, union(set1, set2))
+                uniqueset(set1)
                 gengedcom(set1)
 
 }