[Babase] Latest conversion errors, interpolation proofing
Karl O. Pinc
babase@www.eco.princeton.edu
Tue, 20 Sep 2005 06:11:59 +0000
The latest conversion errors/load script are on papio in:
/scratch/babase/errors
/scratch/babase/loadscript
Please look over the errors (as usual up to the point
where you start seeing program bugs) and investigate.
Following are my notes on the interpolation comparision.
(I'll be sending Cathrine the difference file under separate
cover. Cathrine, when you get it recall that I'm
pretending that the postgres members file has no
origin I, group 9 rows because we add so many
to the new system. Recall that although the diff
sheet points out possible problems you need to
look at the actual data for each instance just to
be sure I've really got the members table right.
Thanks for doing all the grunt work.)
What's with B11? Shouldn't (s)he be dead? Should
(s)he really have a Name or is (s)he a stillbirth etc?
In foxpro the members file seems out of sync with
the census file:
ALE 2003-03-05 2.20 C 0
AMA 2003-03-05 2.20 C 0
Shows up in members but not census (AFICT).
(Unless I'm somehow not getting these census rows
from foxpro???)
You may want to keep a list of foxpro problems that
you find? (E.g. AMI seems to be missing a members
row in foxpro for 2003-08-17.) Maybe not, as
we'll be fixing them anyway?
Crud. I _think_ I just found a bug, so I won't be
sending Cathrine the diff tonight after all. I'll
take a look at it in the morning. If I have to
re-run the conversion I'll eventually replace
the errors and loadscript above, but in the mean
time they're there to be looked at.
Well, it _almost_ looked really good....
Karl <kop@meme.com>
Free Software: "You don't pay back, you pay forward."
-- Robert A. Heinlein