Re: Ngl.contour_map bug when constant field left of 0 lon

Hi Michael,

Dave Brown said that PyNGL can't handle over-wrapping around the globe in this fashion. We realize there are some tools that can do this, but PyNGL doesn't.

However, I did find that 0 to 360 doesn't work either, and it should. This does seem to be an issue due to the constant field.

I've filed a trouble ticket on this (NCL-831 for your reference).

--Mary

On Nov 4, 2010, at 5:21 PM, Mary Haley wrote:

> Sorry Michael, we let this message slip unintentionally.
>
> The first thought that I had on this was that the longitude coordinate array is incorrect or not set at all.
>
> I actually tried running your script when you originally sent it, but it was seg faulting for me. I think my python installation is messed up. I don't think this is an issue with your script.
>
> I'll try to look at this tonight or tomorrow. AGain, sorry for the delay.
>
> --Mary
>
> On Nov 4, 2010, at 7:38 AM, Michael Decker wrote:
>
>> Hi Mary, Dave and everyone,
>>
>> (I'm not really sure if my previous mail made it as it was a little too big with the original eps attachments. So there it is again with pngs instead.)
>>
>> we recently stumbled upon a bug in Ngl.contour_map that happens when there is a constant field value left of 0 degrees longitude on the map. The plot will simply be white/transparent on the whole left side then. As soon as one different value appears in that field, everything works as expected. This only happens left of 0 degrees. If 0 is shifted completely to the left side of the map, then the constant area is displayed just fine (then on the right side).
>>
>> I'm probably not very good at explaining that, so I attached a simple demo script and three output files generated by it (originally eps, converted to png for size).
>> test.png: 0 degrees in the middle of the map, constant field on the left is all white
>> test2.png: left side contains one other value and it works like it should
>> test3.png: 0 degrees shifted to the left of map, same data as for test.png but like this it works just fine
>>
>> Cheers,
>> Michael
>>
>> --
>> Michael Decker
>> Forschungszentrum Jülich
>> Institut für Energie- und Klimaforschung - Troposphäre (IEK-8)
>>
>> E-Mail: m.decker@fz-juelich.de
>>
>> <pyngl_error.py><test.png><test2.png><test3.png>_______________________________________________
>> pyngl-talk mailing list
>> List instructions, subscriber options, unsubscribe:
>> http://mailman.ucar.edu/mailman/listinfo/pyngl-talk
>
> _______________________________________________
> pyngl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/pyngl-talk

_______________________________________________
pyngl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/pyngl-talk
Received on Fri Nov 5 16:35:30 2010