Re: histogram crashes

[Craig Markwardt <craigmnet(at)cow.physics.wisc.edu>]

There have been varying reports in the newsgroup of bugs in HISTOGRAM
over the past year involving versions 5.2-5.3, for Linux or Win NT,
either with or without NaNs.  A quick search of Deja News should find
a lot of these.  I would report it to RSI for sure!

Craig


"R.G.S." <rgs1967@hotmail.com> writes:

> Greetings all,
> 
> I have a situation where histogram is crashing on me, in what seems to be
> a strange manner. (IDL 5.3.1, on WinNT 4 Workstation SP 5)
> 
> Here is info on my data (latitudes):
> LAT             FLOAT     = Array[76, 1624]
> IDL> help,lat(*)
> <Expression>    FLOAT     = Array[123424]
> 
> range of latitudes:      -65.8900      79.9300
> min = :       20.0000
> There are NAN values in the array.
> 
> 
> Here is the offending call to histogram:
> hlat = histogram(lat(*),binsize = float(deltalat), min
> =float(20),REVERSE_INDICES = R,/nan)
> 
> This results in a Norton CrashGuard message and IDL closes.
> 
> 
> Of course,the following call to histogram works with no problems:
> hlat = histogram(lat(*),binsize = long(deltalat), min =float(-1),/nan)
> as does:
> hlat = histogram(lat(*),binsize = float(deltalat), min =float(20),/nan)
> 
> The difference seems to be that a positive "min" crashes and a negative
> "min" is ok when
> the reverse_index keyword is called. For my purposes the reverse_indices
> keyword
> is required.
> 
> Anyone run across this before, and are there any fixes?
> 
> 
> Cheers,
> bob stockwell
> stockwell (at) co-ra.com
> 
> 
> 
> 
> 

-- 
--------------------------------------------------------------------------
Craig B. Markwardt, Ph.D.         EMAIL:    craigmnet@cow.physics.wisc.edu
Astrophysics, IDL, Finance, Derivatives | Remove "net" for better response
--------------------------------------------------------------------------