| cgPlotS problem with NaN values in data [message #88960] |
Tue, 08 July 2014 12:05  |
morganlsilverman
Messages: 46
Registered: February 2013
|
Member |
|
|
Hello,
I'm plotting latitude/longitude data on a map background using cgPlotS and keep getting the error message: % Program caused arithmetic error: Floating illegal operand when it gets to a NaN value in the data. Because I'm plotting this in a resizable graphics window it also brings up the keep command or delete command box each time, eventually crashing my session. Is there a way to get around this without removing the NaN values from my data? I haven't converted the coordinates to metered space because when I did it plotted the lat/lon values in the wrong locations. I think because the data is taken from a model that was run on a map coordinate system maybe. Thanks.
-Morgan
Pro MOZAICPanel3
lonmin = -125
lonmax = -65
latmin = 25
latmax = 60
centerlon = (lonmax-lonmin)/2.0 + lonmin
centerlat = (latmax-latmin)/2.0 + latmin
limit = [latmin, lonmin, latmax, lonmax]
xrange = [lonmin, lonmax]
yrange = [latmin, latmax]
mapCoord = Obj_New('cgmap', 'Stereographic', Ellipsoid='WGS 84', Limit=limit, $
center_lat=centarlat, center_lon=centerlon)
; Set up 2-col by 3-row grid
p = cgLayout([2,3], Aspect=0.65, YGap=3.5, XGap=4, OXMargin=[3,3], OYMargin=[2.5,2.5])
; Draw plots
pos = p(*,5)
restore, './MOZAIC/Trajectories/MOZAIC_Washington_BackTrajectory.sav'
mapCoord->SetProperty, Position=pos, /Draw
cgMap_Grid, map=mapCoord, /box
cgMap_Continents, map=mapCoord, /continents, /countries, /usa
cgplots, lon2500(3,*), lat2500(3,*), map=mapCoord, color='green', thick=4
cgplots, lon3000(3,*), lat3000(3,*), map=mapCoord, color='cyan', thick=4
cgplots, lon3500(3,*), lat3500(3,*), map=mapCoord, color='magenta', thick=4
end
|
|
|
|
| Re: cgPlotS problem with NaN values in data [message #88962 is a reply to message #88960] |
Tue, 08 July 2014 13:18  |
Matthew Argall
Messages: 286
Registered: October 2011
|
Senior Member |
|
|
Try plotting only the points that have non-NaN values.
iGood = where(finite(lon2500[3,*]) and finite(lat2500[3,*]), nGood)
if nGood gt 0 then cgplots, lon2500[3,iGood], lat2500[3,iGood], map=mapCoord, color='green', thick=4
etc.
|
|
|
|
| Re: cgPlotS problem with NaN values in data [message #88963 is a reply to message #88960] |
Tue, 08 July 2014 13:39 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
Morgan Silverman writes:
> I'm plotting latitude/longitude data on a map background using cgPlotS and keep getting the error message: % Program caused arithmetic error: Floating illegal operand when it gets to a NaN value in the data. Because I'm plotting this in a resizable graphics window it also brings up the keep command or delete command box each time, eventually crashing my session. Is there a way to get around this without removing the NaN values from my data? I haven't converted the
coordinates to metered space because when I did it plotted the lat/lon values in the wrong locations. I think because the data is taken from a model that was run on a map coordinate system maybe.
As you point out (somewhat ungraciously by blaming cgPlotS, I might
add), there is a problem when you try to do numbery things with things
that aren't, well, numbers. It's like adding (4 + "atrocious"). What in
the world do you end up with?
But, your problem is not that you can't plot the numbers. To plot the
numbers, you have to first convert them to projected meter space (the
purpose of the map coordinate object you are passing to cgPlotS). This
is the thing that is failing you, I believe.
I don't think you have any choice but to remove the incalculable points.
Where would you locate the darn things on the map otherwise?
Cheers,
David
--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.idlcoyote.com/
Sepore ma de ni thue. ("Perhaps thou speakest truth.")
|
|
|
|
|
|
| Re: cgPlotS problem with NaN values in data [message #88965 is a reply to message #88964] |
Tue, 08 July 2014 15:35 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
Paul van Delst writes:
>
> On 07/08/14 16:39, David Fanning wrote:
>> there is a problem when you try to do numbery things with things
>> that aren't, well, numbers. It's like adding (4 + "atrocious"). What in
>> the world do you end up with?
>
> Ha! I was going to go with the boring old:
>
> Man: "Doctor, it hurts when I do this."
> Doctor: "Well, don't do that!"
>
> but your wordz is bettah.
>
> I'm going to cogitate on how to use the phrases 'numbery things' and '(4
> + "atrocious")' in my near-future interactions with colleagues....
I was worried you were going to say my example was too much like
operator overloading. :-)
Cheers,
David
--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.idlcoyote.com/
Sepore ma de ni thue. ("Perhaps thou speakest truth.")
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
