| Filtering out NaNs [message #8528] |
Wed, 19 March 1997 00:00  |
iarla
Messages: 10
Registered: December 1994
|
Junior Member |
|
|
Hello,
There are some NaNs in my data which keep making my code fall over. I'd
like to either filter them out or avoid them in the calculation, but doing
something like
a(where(a eq !values.f_nan) = missing_flag
doesn't work. Does anyone have any suggestions?
Many thanks,
Iarla.
|
|
|
|
|
|
| Re: Filtering out NaNs [message #8633 is a reply to message #8528] |
Sun, 23 March 1997 00:00  |
Mark Hadfield
Messages: 783
Registered: May 1995
|
Senior Member |
|
|
Stein Vidar Hagfors Haugan <steinhh@rigil.uio.no> wrote in
article <5gp9oq$8hp$1@ratatosk.uio.no>...
>
> NaNs are characterized by the fact that they are *not* equal
> to any number (that's what it says, isn't it :-)
>
> In fact, it's not even equal to itself - and this is the distinguishing
> feature that must be used to pick them out. I.e.,
>
> a(where(a ne a)) = missing_flag
I've always used the finite function, eg:
a(where(1-finite(a))) = missing_flag
However FINITE returns 0 for infinities as well as NaNs and I suppose this
could be a problem in some instances.
I just tried 4 different comparison operators against pairs of NaN's and
found that on my system (IDL 5.0 prerelease on WinNT/Intel) NaN IS equal to
itself, but it's also greater than itself.
IDL> a = !values.f_nan
IDL> print, a eq a, a ne a, a gt a, a lt a
1 0 1 0
% Program caused arithmetic error: Floating illegal operand
Curious.
============================================================ ==
Mark Hadfield NIWA (Taihoro Nukurangi)
PO Box 14-901
m.hadfield@niwa.cri.nz Wellington, New Zealand
|
|
|
|
| Re: Filtering out NaNs [message #8643 is a reply to message #8528] |
Fri, 21 March 1997 00:00 |
Christian Marquardt
Messages: 7
Registered: August 1996
|
Junior Member |
|
|
Peter Webb wrote:
>
> Stein Vidar Hagfors Haugan (steinhh@rigil.uio.no) wrote:
>
> : NaNs are characterized by the fact that they are *not* equal
> : to any number (that's what it says, isn't it :-)
>
> : In fact, it's not even equal to itself - and this is the distinguishing
> : feature that must be used to pick them out. I.e.,
>
> : a(where(a ne a)) = missing_flag
>
> I'd like to hereby nominate Stein for the "Cool Tip of the Week" award!
Well - it will result in an error message from IDL if there are _no_
NaN's or +/-Infinites in that array. Better try something like
idx = where(a ne a) & if idx(0) ne -1 then a(idx) = missing_flag
or
idx = where(finite(a) ne 1) & if idx(0) ne -1 then a(idx) =
missing_flag
I wonder if there's a way to use where without checking if the
condition was met at all, given that you do not know this in
advance.
Regards,
Chris.
------------------------------------------------------------ ------------
Christian Marquardt
Meteorologisches Institut der | tel.: (+49) 30-838-71170
Freien Universitaet Berlin | fax.: (+49) 30-838-71167
Carl-Heinrich-Becker-Weg 6-10 | email:
marq@strat01.met.fu-berlin.de
D-12165 Berlin |
------------------------------------------------------------ ------------
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
