Re: Relational/hierarchical data problem
Date: 20 Feb 2003 03:53:46 -0800
Message-ID: <8792b468.0302200353.694394c6_at_posting.google.com>
Thanks to all who have replied.
Bob - I found what you said very interesting. The number of animals in
this case (goats, incidentally) is likely to be large. There are 41
goats in the current herd and this equates to 208 animal records at
present. The herd will stabilise at about 120 animals and will have a
high turnover (they are meat goats). We are legally obliged to track
the past 5 generations for any given animal, so in a fairly short time
we will be dealing with thousands of animal records.
> It might be possible
I'd agree with that - it seems the best way to retain as much
performance as possible.
According to what my understanding, collateral consanguinity can be
determined by two methods:
Evidently method 2 is a better one to use because there is less
amibiguity.
Bearing the above in mind, I have a question about the consanguinity
values you stated:
> > If parents are half-siblings, one of the grandparents would have
Taking one of the half-siblings as the subject, wouldn't the following
be true?
> to
> > pre-compute the transitive closure in another table then use that to
> > evaluate the shared ancestors.
sibling, sibling, parent, 1
child, parent, parent, 1
child, grandparent, grandparent, 2
nephew, uncle, grandparent, 2
first cousin, first cousin, grandparent, 2
sibling, sibling, parent, (1 + 1) = 2
child, parent, parent, (1 + 0) = 1
child, grandparent, grandparent, (2 + 0) = 2
nephew, uncle, grandparent, (2 + 1) = 3
first cousin, first cousin, grandparent, (2 + 2) = 4
> Consang=2.
- parent -> any grandparent would be Consang=1
- children of either parent -> any grandparent would be Consang=2? I
don't see why one grandparent would have a different degree of
consanguinity to any of the others from the perspective of either
their children or their children's children. How this is affected by
the fact that the parents are half-siblings?
Thanks for your help so far!
Regards
Will Received on Thu Feb 20 2003 - 12:53:46 CET
