comp.lang.idl-pvwave archive: archive » Re: where and NAN numbers

Home » Public Forums » archive » Re: where and NAN numbers

Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend

Re: where and NAN numbers [message #42152]

Wed, 12 January 2005 15:16

Mark Hadfield
Messages: 783
Registered: May 1995

Senior Member

maarten wrote:
> What version are you using, because in 6.1.1. the results are:
> IDL> u = [!values.f_nan, 1, 2, 3]
> IDL> print, where(u eq 3)
> 3
> IDL> print, where(u ge 3)
> 3

The result of a comparison operation between a NaN and an "ordinary"
number will depend on the platform. (At least I vaguely recall some
dicussion of this somewhere, perhaps this group or perhaps another one.)
It is not wise to rely on getting logically consistent results. For
example on IDL 6.1.1 (win32), NaN is not equal to 0 (according to the EQ
and NE operators) but it is neither larger or smaller than 0 (according
to the LT and GT operators). So, as another poster has suggested, it is
good practice to use the FINITE function to establish whether a floating
point number is finite; if it's not the results from comparison
operations will likely be misleading.

--
Mark Hadfield "Ka puwaha te tai nei, Hoea tatou"
m.hadfield@niwa.co.nz
National Institute for Water and Atmospheric Research (NIWA)

Report message to a moderator

[Message index]

		Re: where and NAN numbers By: Mark Hadfield on Wed, 12 January 2005 15:16
		Re: where and NAN numbers By: Haje Korth on Wed, 12 January 2005 06:24
		Re: where and NAN numbers By: maarten on Wed, 12 January 2005 04:08

Previous Topic:	Re: DIALOG_PICKFILE Questions in IDL 6.1
Next Topic:	Re: 3-d visualization (itools while in a program)

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Sat Oct 11 20:53:35 PDT 2025

Total time taken to generate the page: 0.47754 seconds