[ncl-talk] non-alphanumeric variables in ncl-6.3.0

Dennis Shea shea at ucar.edu
Mon Nov 9 11:38:56 MST 2015


Hi Karoline,

Thank you for sending a clean, simple test script and sample files.
This effort is always appreciated by ncl-talk.

The  6.3.1 version has addressed the issue.

On Fri, Nov 6, 2015 at 8:33 AM, Dennis Shea <shea at ucar.edu> wrote:
> Hello,
>
> Could you please send us a clean version of your script and a sample data file?
>
> ftp ftp.cgd.ucar.edu
> anonymous
> email
> cd incoming
> put ...
> put ...
> quit
>
> Then, after successful transfer, send us the names of the files you transferred.
>
> ---
> If you can make them available via the WWW that would be fine also.
>
>
> THX
> D
>
> On Fri, Nov 6, 2015 at 8:02 AM, kblock <karoline.block at uni-leipzig.de> wrote:
>> Dear all,
>>
>> I have encountered a problem in ncl version 6.3.0 in reading
>> non-alphanumeric variables from a net-cdf file (File format : netCDF4
>> ZIP), which is not occuring in ncl version 6.0.0.
>> I have tested it using the same script and the same data on different
>> servers with these two ncl version installed.
>>
>> In the documentation it says you should read in like this:
>> x = f->$"ice cream+oreo-cookies...yummy!"$
>>
>> But this only works in the older version!
>> For ncl 6.3.0 it is advisable to change the variable name before reading
>> in the file (e.g. with cdo chname,oldvar,newvar ifile ofile).
>> Otherwise it seems that something similar as in asciiread will happen,
>> that for the float and double type variables , the string 'nan' is read
>> as a valid numerical value.
>>
>> With ncl-6.3.0, the single-dimensioned float variable "/var0.4" was
>> mostly read in as "-nan",
>> while with ncl-6.0.0 the same variable (now without the "/",even though
>> it didn't matter for the reading in) was read in perfectly fine with the
>> same command:
>> x = f->$"var0.4"$
>>
>> After I changed the variable name to "var04" both versions work just fine!
>>
>> I like to encourage the NCL Team to look at this and hopefully fix this
>> problem!
>> Users are strongly advised to keep this in mind when reading data!
>>
>> Cheers,
>> Karoline
>>
>>
>> _______________________________________________
>> ncl-talk mailing list
>> ncl-talk at ucar.edu
>> List instructions, subscriber options, unsubscribe:
>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk


More information about the ncl-talk mailing list