branches - duplicate persons in the list

2010-01-09
2013-05-30
  • Paul Seesink

    Paul Seesink - 2010-01-09

    Running the branch-list on a certain family name (Vogel), I see that a number of persons occurs twice. I insert here one example, compare the entries 16.1 and 20, that I include here manually in this topic :
    16. GERHARD VOGEL
    1. HEINRICH VOGEL  *1843 MARRIED: 1875 ELISABETH NIENHUIS *1856

    20. HEINRICH VOGEL  *1843 MARRIED: 1875 ELISABETH NIENHUIS *1856

    The result is that I get from a certain Gedcom a numbered list of some 50 entries (most of these of course with a number of descendents underneath), But without these duplicates the numbered list would have had only some 40 entries, because I have some 40 (yet) unrelated branches under that family name.
    I would have expected the last case, when looking in the branch list.

    Maybe there is another explanation, but I can only imagine that the list is produced by an algorithm like:
    1. find the first person within the selected branch resulting in entry 1, with descendants, if any.
    2. find the next person, that has not yet been outputted in previous entries.
    In that algorithm a person can occur twice: 1st as a new entry, 2nd as a descendant under another person
    Very annoying where it is the intention to get insight in the number of independent branches in the selected family name.

    An algorithm to get the last case (no duplicates), could be quite easily realized, by:
    • first finding for a person its very first forefather and only then producing, starting with its forefather, a new entry in the branch-list.

     
  • Anonymous - 2010-01-12

    Can I look at this on your site to get a better idea of the issue, as I can't find any way to reproduce it on any of my test installations.

     
  • Paul Seesink

    Paul Seesink - 2010-01-16

    Kiwi_pgv answered me: Can I look at this on your site to get a better idea of the issue, as I can't find any way to reproduce it on any of my test installations.
    Yes, of course. It occurs in Gedcom "Vogel" in http://www.seesink.relatedfamilies.com/index.php?ctype=gedcom
    Is this enough for you to access?

     
  • Anonymous - 2010-01-17

    Yes, that is exactly what was needed, and as I suspected, there is nothing wrong with the code. The problem is in your data. If you look at the raw GEDCOM data for I1991 (HEINRICH VOGEL) he has no FAMC record - so as an individual he has no parent family. That is why he appears as Branch 20.

    But if you look at the family record of his father (F3332) it DOES list Heinrich as a child, and so he appears as such (branch 16.1). It would be wrong to spend a lot of time writing code to ignore data that has been entered incorrectly IMHO. How did you create this GEDCOM, manually or by import from some other software? It certainly isn't a problem you would get if entering data in PGV itself in the normal way.

    Cross references linking children to families must ALWAYS work both ways. The INDI record must include "1 FAMC @F3322@" and the family record must include "1 CHIL  @I1991@"

    If you tidy this up, then the Branches list will be able to function normally.

    If there are other similar "duplicates" you can be sure they have the same issue.

    I recommend you go to the GEDCOM configuration page on your site, and against each GEDCOM file click on "Check". Leave it set to "Critical" errors only, and it will list all such examples of missing cross-references for you.

     
  • Paul Seesink

    Paul Seesink - 2010-01-17

    This Gedcom Vogel is indeed import from some other software, from which I moved all my data tto PhpGedView.
    As you explained me, I am left with some inconsistent data, that I needed to make consistent.
    I want to solve that and with the directions you gave I expect I can do so.
    Note: this Gedcom was created deliberatedly from a collection of small Gedcom's of the "familysearch international genealogical index" and I was already busy in my former software in the merging process of persons.
    Thanks.

     

Log in to post a comment.