Hi Russell,
> How does "2003-07-21 18:00 +0000" cover "2003-07-21 18:25 +0000"? Is it
> because the time from the 18:25 sample to the 18:00 sample is less than
> half-way from 18:25 to 17:25? (I hope that's clear.)
Essentially yes. The actual logic is in the Set.merge1DSets()
method. If the values of the merged Set are passed through
the doubleToIndex() method of each contributing Set, they
produce all the indices of that contributing Set.
> Also, the program that this was extracted from does use setSet. I've
> just tried modifying the test program by making a new Gridded1DDoubleSet
> with all the time samples in order and passing that to setSet(), and
> those two samples are still missed.
> . . .
> setSet() did not seem to have the desired effect. Have I missed
> something?
You've uncovered a bug. setSet() would probably work if you
called it after the data were displayed the first time, so
the auto-scaling wouldn't override ride it. But an explicit
call to setSet() should disable future calls by auto-scaling,
as with ScalarMap.setRange(). The fix is in the jar files at:
ftp://ftp.ssec.wisc.edu/pub/visad-2.0/untested/
Cheers,
Bill
