[Babase] Re: repstats rebuild error
Niki Learn
nlearn at princeton.edu
Fri Mar 5 12:19:31 EST 2010
On 03/05/2010 10:54:21 AM, Niki Learn wrote:
> Okay, strike that. I did do something to PAT - I censored her.
What exactly does that mean? Did you change the Statdate? From what
to what?
HER PREVIOUS STATDATE WAS 11 JUNE 1993, which matches the last day she
appears in the census record. Then she was listed as alive of course, but
now she is censored to make her dead to babase but indicate that we don't
actually know anything about if or when she died.
I DID CHANGE THE STATDATE but I moved it farther ahead in time to the 24th
to match a wounds record, which I cross referenced while doing this to see
which of these baboons might really be dead. So it seems like the Cycgap
records stretch past the statdates if the baboons are alive?
SO REBUILDING cycgapdays would fix the problem? That sounds reasonable to
me under the circumstances. It's not every day we change statdates and
statuses around...
> So
> this is
> another weird fruity error like those neighbors ones that is popping
> up from
> that. How do these cycgaps tables work and interact with the state &
> statdate?
It's supposed to do the right thing automagically. But the code
has not ever had a lot of work to do because nobody's changed many
statdates that old. The Status should not matter, perhaps
excepting "alive". The Statdate is what counts. At least that
what I think now. I need to look closely to figure out what's
supposed to happen and what went wrong.
>
> This could have occurred with a number of other females too since the
> females that got censored were all from groups that we stopped
> monitoring at
> some point and they were usually monitored sort of sporadically for a
> while
> before all monitoring stopped so all of those females are probably in
> cycgaps...
Best case: I fix the problem and get you to re-do the changes in
babase_test and re-build REPSTATS. That way I'm sure that my fix
works with everything. We may decide that's overkill and just
solve the problem, rebuild CYCGAPDAYS to fix it, and move on.
FYI
CYCGAPS itself is probably not "right", in that nobody's looked
closely at just when observations really started and stopped.
We just poked at it with a stick during the conversion process
to ensure that there were no inconsistencies with the existent data.
This does not mean that what's in CYCGAPS reflects what really
happened and is consistent with data that's not been entered.
Karl <kop at meme.com>
Free Software: "You don't pay back, you pay forward."
-- Robert A. Heinlein
_______________________________________________
Babase mailing list
Babase at www.eco.princeton.edu
http://www.eco.princeton.edu/mailman/listinfo/babase
More information about the Babase
mailing list