[Babase] CYCSTATS Proofing
Catherine Markham
amarkham at princeton.edu
Fri Jun 30 14:35:58 EDT 2006
Karl,
Thanks for your input earlier on the repro tables. Notes and sample
documentation to include for the issue of auto mdates added after a
female's statdate are below...
Catherine
> 5. I did a check similar to the one above but with the automated
> mdates. On all observed mdates, if the female was in CYCSTATS, she was
> either in state "M" or "S". That part is good. However, there were
> several females that were matured but had no CYCSTATS record on an
> mdate. For all of these, the mdate in question was the "dangling" auto
> mdate we've been talking about recently (the auto mdate is calculated
> from the last observed ddate and has no corresponding tdate or ddate).
> In these cases, the last CYCSTATS rows was on the female's statdate
> (state was always "D" - good).
>
> Questions for Karl: (1) Does the logic of how this is working match
> your expectations? If so, I think we need to specify what is going on
> in the documentation (it took me a while querying to figure out what the
> rules are). Let me know if it is already there - perhaps I missed it
> any maybe we just need to make it more explicit? (2) Is it going to be
> an issue that there are dates in MTDCYCLES (and also CYCPOINTS) that are
> after a female's statdate?
If I understood correctly, Karl said it probably will not cause database
errors to have auto mdates after the statdate SO LONG AS the statdate is
not a deathdate (he is going to investigate subset of records where the
female is dead on the statdate). For the other records (the ones
involving living females), we agreed that a note should be added to the
documentation to alert users as to what happens with CYCSTATS in these
situations. Here's a first stab the writing of that text:
"Unless reproductive monitoring was affected by gaps in observation,
CYCSTATS records use all available CYCPOINTS data and continue through a
female's statdate. One notable exception exists in the case of
automatically generated mdates, which potentially occur after a female's
statdate. If an automatically generated mdate is after a female's
statdate, there will be no corresponding cycle data in CYCSTATS beyond
the statdate (i.e., CYCSTATS will not directly reflect the data in
CYCPOINTS for days later than BIOGRAPH.Statdate)."
Karl, we talked about this as directly related to CYCSTATS - should
comparable text be added to REPSTATS? Also, do you think we need to
make the distinction between statdates and deathdates in the
documentation? I'm thinking not because it sounded like you were going
to get rid of the auto mdates after a deathdate, so these will no longer
be in CYCPOINTS anyway.
Catherine
More information about the Babase
mailing list