| Re: Unsigned Integer Math Problem [message #70527] |
Wed, 21 April 2010 10:19 |
jeanh
Messages: 79
Registered: November 2009
|
Member |
|
|
> Clearly, I don't want the minimum x range value to be less
> than zero in this case, but I also don't want to force the
> value to be zero if the minimum I want is somewhat higher
> than this, say 1200. How do I test for this? Clearly, this
> does not work:
>
> min_xrange = (omin - binsize)> 0
>
> Since this number 64810 *is* larger than zero, and WRONG!
what about doing if (omin-binsize) gt omin then min_xrange=0
Jean
|
|
|
|
| Re: Unsigned Integer Math Problem [message #70528 is a reply to message #70527] |
Wed, 21 April 2010 07:46  |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
Craig Markwardt writes:
> Hi David, I think there is no way to do exactly what you want. But
> this is a "doctor it hurts when I do this" problem. For plotting
> purposes, there is absolutely no harm in upcasting to a signed
> integer, and *then* doing your devious arithmetic.
>
> For that matter, why even bother with integers? For plotting
> purposes, make your calculations in double precision!
Well, as I just discovered in another program where this
same thing occurs, this is exactly what I do. :-(
I brought this up mostly as a cautionary tale about
doing math with unsigned integers. This is the second
time in a week unsigned integers have given me
heartburn.
Cheers,
David
--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Sepore ma de ni thui. ("Perhaps thou speakest truth.")
|
|
|
|
|
|
| Re: Unsigned Integer Math Problem [message #70530 is a reply to message #70529] |
Wed, 21 April 2010 07:34 |
Craig Markwardt
Messages: 1869
Registered: November 1996
|
Senior Member |
|
|
On Apr 21, 10:12 am, David Fanning <n...@dfanning.com> wrote:
> Folks,
>
> I've run into a problem with my Histoplot code this morning.
>
> It is extremely important to the Histogram command that the
> data type of the BINSIZE argument be the same as the data
> type of the data for which the histogram is being calculated.
> I don't know why this is the case, but it is.
>
> In any case, I'm extremely careful about this. But this
> is giving me a problem when I try to make a histogram plot
> of an image that is stored as unsigned integers (UINT).
>
> Basically, to make my plot I take the output minimum
> and maximum from the histogram command and subtract (or
> add) a full binsize to those numbers to give the X axis
> range of the plot.
>
> My problem is this. The OMIN of the histogram is 0, the
> binsize is 726.
>
> IDL> help, omin, binsize
> OMIN UINT = 0
> BINSIZE UINT = 726
>
> When I make the calculation for the minimum data range
> of my axis, I do this:
>
> IDL> min_xrange = omin - binsize
> IDL> Help, min_xrange
> MIN_XRANGE UINT = 64810
>
> Now, this causes the minimum x range to be larger than the
> maximum x range and results in complete chaos downstream.
>
> Clearly, I don't want the minimum x range value to be less
> than zero in this case, but I also don't want to force the
> value to be zero if the minimum I want is somewhat higher
> than this, say 1200. How do I test for this? Clearly, this
> does not work:
>
> min_xrange = (omin - binsize) > 0
>
> Since this number 64810 *is* larger than zero, and WRONG!
>
> I guess my real question is this: How do I do arithmetic
> operations with unsigned integers in a way that preserves
> the nature of unsigned integers?
>
> Any ideas on this?
Hi David, I think there is no way to do exactly what you want. But
this is a "doctor it hurts when I do this" problem. For plotting
purposes, there is absolutely no harm in upcasting to a signed
integer, and *then* doing your devious arithmetic.
For that matter, why even bother with integers? For plotting
purposes, make your calculations in double precision!
Craig
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
