[ncl-talk] which value has dmsg

[Dennis Shea]

not )FillValue .... _FillValue

On Mon, May 2, 2016 at 7:40 AM, Dennis Shea <shea at ucar.edu> wrote:

> The attribute )FillValue associated with the input variable is the 'dmsg'.
> This is assigned in the C-interface between NCL and the fortran EOF
> subroutine .
>
> On Fri, Apr 29, 2016 at 5:37 PM, Erik J. Schaffernicht <
> eschaffe at uni-koeln.de> wrote:
>
>> In the ncl eof fortran code, there is a variable    dmsg
>>
>> How does it get its value?     How is the value transferred from the
>> NCLscript  to fortran?
>> Is dmsg taken from the metadata, i.e.   @missing_value  ressource of the
>> array I pass to eofunc()  ?
>>
>>
>>
>> Can I run an eof, where I set 75% of the grid points to missing_value
>>  in my input array?
>>
>> If yes, I expect my eof outcome to be the same if I just give as input,
>> i.e.:
>>
>> VersionA
>> Say 10 grid points, that have everywhere valid data and that are concat.
>> to one array.
>>
>> VersionB
>> Say 40 grid points, the 10 of VersionA concatenated with 30 other.
>> The 30 other g. p. consist of missing_value for all time steps, so only
>> the 10 grid points equal to those of VersionA  contain data for every time
>> step.
>> The eof output will be the same as for VersionA, right?
>>
>> Or should I set these 30 grid points equal to value zero,  =0, instead of
>> to   missing_value?
>>
>> My aim is to pass a fixed dimension size array to eof and
>> mask/set-to-zero/ignore  all grid points that I do not need during that eof
>> calculation.
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20160502/fe28a664/attachment.html