Hi John,
I guess there is a reason for modelers to use this missing value for vmax,
vmin and valid range. As it is quite expansive to calculate each model
output (snapshot) on all variables' max, min and range (we are talking about
O(N), and they use the missing value for max,min and range by default for
undefined. Robert did a bit change on his Panoply and it worked fine just
like Ferret/GrADS to read such a dataset.
I just wonder for better/more robust NJ, you may consider this range,
vmax and vmin (when defined by using missing_value) just as same as those
missing_value at some grid for all variables --- instead of treat them as
real sense of max and min.
I have not yet heard back from data producers, but I guess that might be the
reason for having this kind attributes.
On Thu, Oct 28, 2010 at 8:51 PM, Robert B. Schmunk <
Robert.B.Schmunk@xxxxxxxx> wrote:
>
> John,
>
> I have been looking at Ying's data this evening, and I'm a bit
> confused about the metadata that are reported for the variables.
>
> I had been looking at a different file, one downloaded to disk.
> but same source and same problem.
>
> The Variable.toString() and VariableDS.toStringDebug() methods
> report for an example variable PS that there are attributes:
>
> float PS(time=1, lat=721, lon=1152);
> :long_name = "surface_pressure";
> :units = "Pa";
> :fmissing_value = 9.9999999E14f; // float
> :standard_name = "surface_pressure";
> :vmin = 9.9999999E14f; // float
> :vmax = 9.9999999E14f; // float
> :_CoordinateAxisType = "Pressure";
>
> Calling Variable.getAttributes() reports the same list of attributes
> and no others.
>
> Since "fmissing_value", "vmin" and "vmax" are not recognized
> attributes, I figured they would be ignored and the data would
> be plotted. Instead, blank plot.
>
> Then I took look at what ncdump had to say about the same variable
> and it reported:
>
> float PS(time, lat, lon) ;
> PS:long_name = "surface_pressure" ;
> PS:units = "Pa" ;
> PS:_FillValue = 1.e+15f ;
> PS:missing_value = 1.e+15f ;
> PS:fmissing_value = 1.e+15f ;
> PS:scale_factor = 1.f ;
> PS:add_offset = 0.f ;
> PS:standard_name = "surface_pressure" ;
> PS:vmin = 1.e+15f ;
> PS:vmax = 1.e+15f ;
> PS:valid_range = 1.e+15f, 1.e+15f ;
>
> And this left me completely mystified. Where did the other attributes
> come from, and how is that NJ seems to be using them without reporting
> their existence? Anything you can do to clear the fog would be appreciated.
>
> rbs
>
>
>
>
>
> On Oct 28, 2010, at 20:14, Ying at NCCS wrote:
>
> > John,
> >
> > Thank you! This data was produced by NASA GMAO at GSFC. I will contact
> > them to find out the reason to have this range
> >
> > valid_range = 9.9999999E14f, 9.9999999E14f; // float --- which does not
> > seem to make sense at all
> >
> > it is funny though, Ferret is a lot more liberal (maybe GrADS as well) to
> > put up with this.
> >
> >
> > On Thu, Oct 28, 2010 at 7:35 PM, John Caron <caron@xxxxxxxxxxxxxxxx>
> wrote:
> >
> >> On 10/28/2010 4:17 PM, Ying at NCCS wrote:
> >>
> >> Hi John,
> >>
> >> I run into this problem using netcdf-java toolsUI to open either local
> >> netcdf-4 file or via opendap, here is an example link
> >>
> >>
> >>
> http://esg.nccs.nasa.gov/thredds/catalog/ying/readonly/catalog.html?dataset=ying/readonly/e562p1_fp.inst3_3d_asm_Nv.20100907_00z%2B20100909_1200z.nc4
> >>
> >> it just shows all -Infinity on the color-bar.
> >>
> >> I was first having trouble to use NASA GISS Panoply (which uses
> >> netcdf-java), then I tried this recent version of netcdf-java 42.
> ToolsUI
> >> (downloaded from UNIDATA) --- it can read netcdf-4 file (the above
> link)
> >> but it shows same max and min values of each variable. As a result, it
> just
> >> gives a blank plot.
> >>
> >> Then I tried toolsUI-4.2.jar and it shows the same problem.
> >>
> >>
> >> ftp://ftp.unidata.ucar.edu/pub/netcdf-java/v4.2/toolsUI-4.2.jar
> >>
> >>
> >> by the way, I can use Ferret (most recent version) to open this netCDF-4
> >> file and makes right plots.
> >>
> >>
> >>
> >> _______________________________________________
> >> netcdf-java mailing listnetcdf-java@xxxxxxxxxxxxxxxx
> >> For list information or to unsubscribe, visit:
> http://www.unidata.ucar.edu/mailing_lists/
> >>
> >>
> >> Hi Ying:
> >>
> >> that file has this attribute on the data variables:
> >>
> >> :
> >> this is a mistake, it says that there are no valid values. if it can be
> >> removed from the file, it will work. can you contact the file author to
> have
> >> it removed?
> >>
> >> looking at a random variable:
> >>
> >> float SLP(time=1, lat=721, lon=1152);
> >> :_CoordinateAxes = "time lat lon ";
> >> :long_name = "sea_level_pressure";
> >> :units = "Pa";
> >> :_FillValue = 9.9999999E14f; // float
> >> :missing_value = 9.9999999E14f; // float
> >> :fmissing_value = 9.9999999E14f; // float
> >> :scale_factor = 1.0f; // float
> >> :add_offset = 0.0f; // float
> >> :standard_name = "sea_level_pressure";
> >> :vmin = 9.9999999E14f; // float
> >> :vmax = 9.9999999E14f; // float
> >> :valid_range = 9.9999999E14f, 9.9999999E14f; // float
> >>
> >> i would say that it would be better to have:
> >>
> >> float SLP(time=1, lat=721, lon=1152);
> >> :long_name = "sea_level_pressure";
> >> :units = "Pa";
> >> :_FillValue = 9.9999999E14f; // float
> >> :fmissing_value = 9.9999999E14f; // float
> >> :standard_name = "sea_level_pressure";
> >> :vmin = 9.9999999E14f; // float
> >> :vmax = 9.9999999E14f; // float
> >>
> >> John
> >>
> >>
> >> _______________________________________________
> >> netcdf-java mailing list
> >> netcdf-java@xxxxxxxxxxxxxxxx
> >> For list information or to unsubscribe, visit:
> >> http://www.unidata.ucar.edu/mailing_lists/
> >>
> > _______________________________________________
> > netcdf-java mailing list
> > netcdf-java@xxxxxxxxxxxxxxxx
> > For list information or to unsubscribe, visit:
> http://www.unidata.ucar.edu/mailing_lists/
>
> --
> Robert B. Schmunk, Robert.B.Schmunk@xxxxxxxx
> NASA Goddard Institute for Space Studies, 2880 Broadway, New York, NY 10025
>
>
>
