Re: Problem writing masked array to netcdf

I think I may have figured out what is happening here.

I've noticed in a few of my scripts, I have to close and re-open the grib2
file prior to reading new messages. I'm not sure why this is happening, or
why a fix like grib2fh.seek(0) doesn't work. When I applied that fix to
this code, the data reads cleanly and writes cleanly to the file.

This isn't the first issue I've noticed with the file pointer in a grib2
file that was opened using PyNio, so you guys might have a bug. I'm using
version 1.3.0b5.

Dan

On Fri, Mar 23, 2012 at 10:06 AM, Daniel Gilmore <daniel.gilmore@noaa.gov>wrote:

> Good morning.
>
> (I originally wrote this and it ended up mixed in with another thread, so
> I'm breaking it out into its own message. Sorry about the mix up.)
>
> I'm having an issue writing masked data to a netcdf file. I've read this
> section <http://www.pyngl.ucar.edu/bugs.shtml#get_assign_value> and
> applied the work-around to my code. The code to read and write the data
> looks like this:
>
> v = self.grib2fh.variables[grib2var]
> mv = v[:].data
> self.ncfh.variables[cdfvar].assign_value(mv)
>
> ...where grib2var is the grib2 variable name and cdfvar the corresponding
> cf-compliant variable name.
>
> When I read the value from the grib2 file, the data retrieved appears
> correct, but when I write to and then retrieve the data from the netcdf
> file, the data is all fill values and none of the data looks like it was
> written. I'm including the arrays for both below.
>
> Any thoughts?
>
> --------------------------------------------------------------
>
> (Pdb) self.grib2fh.variables[grib2var][:].data[0][186]
> array([ 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> - - - repeated missing value data omitted for bevity - - -
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 2.84769989e+02,
> 2.84540009e+02, 2.84299988e+02, 2.89919983e+02,
> 2.90049988e+02, 2.90209991e+02, 2.90389984e+02,
> 2.90579987e+02, 2.90769989e+02, 2.92559998e+02,
> 2.92540009e+02, 2.92479980e+02, 2.91359985e+02,
> 2.91429993e+02, 2.91440002e+02, 2.91399994e+02,
> 2.91279999e+02, 2.87149994e+02, 2.86579987e+02,
> 2.86059998e+02, 2.85669983e+02, 2.85399994e+02,
> 2.85299988e+02, 2.85320007e+02, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20], dtype=float32)
> (Pdb) self.ncfh.variables[cdfvar][:].data[0][186]
> array([ 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> - - - repeated missing value data omitted for brevity - - -
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20, 1.00000002e+20,
> 1.00000002e+20, 1.00000002e+20], dtype=float32)
>
> --
> --
> Daniel Gilmore (AceInfo Solutions)
> PGB/MDL/OST/NWS/NOAA
> Office #: SSMC2-10357
> Phone: 301.713.0224 x101
> Cell: 703.232.9603
>

-- 
-- 
Daniel Gilmore (AceInfo Solutions)
PGB/MDL/OST/NWS/NOAA
Office #: SSMC2-10357
Phone: 301.713.0224 x101
Cell: 703.232.9603

_______________________________________________
pyngl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/pyngl-talk
Received on Mon Mar 26 11:09:10 2012