[Babase] repstats Q for Karl
kfenn
kfenn at princeton.edu
Fri Jun 19 15:18:01 EDT 2009
Karl O. Pinc wrote:
>
> On 01/15/2009 12:27:42 PM, kfenn wrote:
>> Hi Karl,
>>
>> We have a lady in Repstats (LADHA) who remains in a pregnant state
>> for up to 394 days (dins) in 2007 & 2008. She gave birth on 28 May
>> 2007 but continues to show up as "P" in the repstats table after this
>> date. The birth occurred on day 179 and somewhere I recall that we
>> have a rule that includes a 180 day gestation period. Of course this
>> sounds nuts because obviously we have stillbirths and aborts before
>> 180 days.
>>
>> Do you have any guesses why she might remain in this state so long?
>> Why didn't the birth trigger a change in state in repstats? I
>> checked the documentation but I can't find the infosurrounding the
>> triggers for change in state.
>
> The basic problem is that the triggers are not updating repstats
> or cycstats. Theses must be manually updated by running the
> build program whenever any reproductive info is changed.
>
> My guess is that the rebuild_all_repstats program was run
> after all the sexual cycle information was entered
> but before the birth was entered.
> LAD does not have another cycle until 2008-06-06,
> so as far as the system is concerned she's still pregnant
> until then. Of course after the birth is entered the
> next time rebuild_all_repstats is run the repstats
> file is fixed -- as we see in the table now.
>
I checked my logs on the demography update. LAD7 (LAQ) was an 07a
birth. She was on the births .txt file that I archived after uploading
to Babase. However, she was a quirky one that required her pid to be
omitted because, I think, the conception cycle also was getting uploaded
in 07a.
I guessed that omitting her pid would make the repstats table keep her
in a constant pregnant state. I tested this in babase_test by deleting
the pid for another individual, NAP8, and rebuilding REPSTATS. Sure
enough, NAP never transitions after the birth of her offspring when the
pid isn't in BIOGRAPH. Makes perfect sense since that is the link
REPSTATS has to BIOGRAPH.
So the question is...did I really forget to update her BIOGRAPH record
with a pid at the end of 07a and not notice right through 2009? During
the babase_test runs for 07a I have notes on updating her pid but my
notes on the final Babase run gets pretty bogged down in troubleshooting
demog_census bugs (07a was the first update we ever did so it wouldn't
be too surprising if that change slipped through the cracks when I got
distracted with bigger problems). I don't recall what was in Babase
versus babase_test when Laurence first brought this to my attention in
January. But I don't remember having an 'A-Ha' moment when I realized
the pid was missing.
I've got lots of clear records that REPSTATS has been rebuild at the end
of every demog update. Indeed the table doesn't exist unless I rebuild
because we HAVE to delete it at the beginning of every update. Data
don't go in unless we clear REPSTATS, CYCSTATS, MMINTERVALS, MDINTERVALS.
So, for the record, I'm going to assume I forgot the pid, realized the
error, and plugged it in at some point in 2008 before the 08b update
occurred. I think REPSTATS got updated correctly after the 08b
update. I'm pretty sure the pid was in there when Laurence brought the
error to my attention. Although I can't exactly follow the paper trail
to say this happened...all thing being equal it's the most logical and
simplest explanation since 07a was such a chaotic update. I don't think
we have to worry about the REPSTATS programming being wrong.
Tabby
--
Tabby Fenn
Research Assistant
Dept of Ecology and Evolutionary Biology
401 Guyot Hall
Princeton University
Princeton, NJ 08544
609 258-6898 (Ph)
609 258-2712 (Fx)
More information about the Babase
mailing list