[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[netCDF #NFO-217382]: Variables not displayed with ncdump



> Here is the link to my file ftp://ftp.vsg3d.com/outgoing/tmp/O3.nc
> Please take note that this is a file for debug only not for use.

Thanks, I've looked at that file and manually parsed the format bytes
to verify that the header shows only one variable named "O3".  There
is no room in the header for any other variables, as the floating
point values of O3 start right after that variable definition, at
offset 1e34(hex) = 7732 bytes from the begining of the file.  The
output of ncdump is correct in this case, that the file has only a
single variable named "O3".

I think the other "variables" Panoply finds are not netCDF variables
in the file, but rather synthesized variables created by Panoply to
provide a coordinate system to the file, which is lacking a
CF-compliant coordinate system.  I've looked at the Panoply
documentation to see if this is explained there, but can't find an
explanation of how these extra synthesixed variables are generated
from a file that lacks a CF-compliant coordinate system.

I suggest you contact the Panoply developer or support email list to 
verify that Panoply generates psuedo-variables not in the input file.

--Russ

> 2010/11/15 Unidata netCDF Support <address@hidden>
> 
> > Hi,
> >
> > > I know that coordinate variables must have the same name than dimensions.
> > I
> > > also have some files with variables that defines the attribute
> > 'coordinates'
> > > in order to find the coordinate variables if the names do not comply the
> > > first case. And in another file, the attributes 'positions' and
> > > 'connections' are declared in order to find such coordinates. For these
> > > files, there's no problem, all variables are found. However, for this one
> > > file, there is no easy way to find such variables as they are not
> > > recognized.
> > >
> > > What I don't understand is how Panoply can display the main variable with
> > > the right coordinates whereas ncdump cannot even find them. I saw that
> > > Panoply was using ncBrowse to read the netcdf file, but if I use it, the
> > > variables are not found.
> >
> > Panoply uses the netCDF-Java API rather than the C API which ncdump uses.
> >  Still
> > that doesn't explain why it should see variables ncdump can't, because the
> > netCDF
> > format is API-neutral.  Can you make the file in question available to us,
> > either as
> > an email attachment or on a server, so that we can investigate further?
> >  Thanks!
> >
> > --Russ
> > >
> > > 2010/11/15 Unidata netCDF Support <address@hidden>
> > >
> > > > Hi,
> > > >
> > > > > When I run "ncdump -k ..." it returns "classic" format. And the
> > netcdf
> > > > > library that I uses is the 4.1.1 version.
> > > >
> > > > ncdump -h should show all the variables names shapes and types, but not
> > the
> > > > values
> > > > of any variables.  ncdump -c shows the information ncdump -h shows plus
> > the
> > > > values
> > > > of coordinate variables, but it looks like your netCDF file has no
> > > > coordinate variables.
> > > >
> > > > A netCDF coordinate variable has the same name as the corresponding
> > netCDF
> > > > dimension.
> > > >
> > > > --Russ
> > > >
> > > > > This is the dump that I obtain in Panoply if I select "Only plottable
> > > > > variables" (or using ncdump) :
> > > > >
> > > > > dimensions:
> > > > > TSTEP = UNLIMITED;   // (1 currently)
> > > > > LAY = 15;
> > > > > ROW = 398;
> > > > > COL = 478;
> > > > > variables:
> > > > > float O3(TSTEP, LAY, ROW, COL);
> > > > > O3:long_name = "O3              ";
> > > > > O3:units = "ppmV            ";
> > > > > O3:var_desc = "Variable O3  ";
> > > > >
> > > > > Otherwise, if I select "All variables", 4 other variables are
> > displayed:
> > > > > double x(COL=478);
> > > > > :units = "km";
> > > > > :long_name = "synthesized coordinate from XORIG XCELL global
> > > > > attributes";
> > > > > :_CoordinateAxisType = "GeoX";
> > > > > double y(ROW=398);
> > > > > :units = "km";
> > > > > :long_name = "synthesized coordinate from YORIG YCELL global
> > > > > attributes";
> > > > > :_CoordinateAxisType = "GeoY";
> > > > > double level(LAY=15);
> > > > > :units = "sigma";
> > > > > :long_name = "synthesized coordinate from VGLVLS global attributes";
> > > > > :positive = "down";
> > > > > :_CoordinateAxisType = "GeoZ";
> > > > > int time(TSTEP=1);
> > > > > :units = "seconds since 2010-10-27 00:00:00 UTC";
> > > > > :long_name = "synthesized time coordinate";
> > > > > :_CoordinateAxisType = "Time";
> > > > >
> > > > > 2010/11/11 Unidata netCDF Support <address@hidden>
> > > > >
> > > > > > Hi,
> > > > > >
> > > > > > > When I use ncdump -h in order to display the header of my netcdf
> > > > file,
> > > > > > > there is only one variable displayed. However, if I use Panoply
> > and
> > > > > > > choose to display 'all variables', the missing ones are also
> > > > displayed.
> > > > > > >
> > > > > > > Can someone explain why some variables are not displayed and
> > > > recognized
> > > > > > > using ncdump ?
> > > > > >
> > > > > > No, that's puzzling.  Can you run "ncdump -k ..." on your netCDF
> > file
> > > > > > to see what kind of format ncdump thinks it is?  If ncdump
> > complains
> > > > > > that the "-k" option is not valid, it's possible you are using an
> > > > > > older version of ncdump that cannot access netCDF-4 files, but that
> > > > > > you have a version of Panoply which understands netCDF-4 files.
> >  But
> > > > > > in that case, it should not be able to display any variables.
> > > > > >
> > > > > > Can you make your netCDF file available to us via FTP or HTTP so we
> > > > > > can try to replicate the problem here?
> > > > > >
> > > > > > --Russ
> > > > > >
> > > > > > Russ Rew                                         UCAR Unidata
> > Program
> > > > > > address@hidden
> > http://www.unidata.ucar.edu
> > > > > >
> > > > > >
> > > > > >
> > > > > > Ticket Details
> > > > > > ===================
> > > > > > Ticket ID: NFO-217382
> > > > > > Department: Support netCDF
> > > > > > Priority: Normal
> > > > > > Status: Closed
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > >
> > > > Russ Rew                                         UCAR Unidata Program
> > > > address@hidden                      http://www.unidata.ucar.edu
> > > >
> > > >
> > > >
> > > > Ticket Details
> > > > ===================
> > > > Ticket ID: NFO-217382
> > > > Department: Support netCDF
> > > > Priority: Normal
> > > > Status: Closed
> > > >
> > > >
> > >
> > >
> >
> > Russ Rew                                         UCAR Unidata Program
> > address@hidden                      http://www.unidata.ucar.edu
> >
> >
> >
> > Ticket Details
> > ===================
> > Ticket ID: NFO-217382
> > Department: Support netCDF
> > Priority: Normal
> > Status: Closed
> >
> >
> 
> 

Russ Rew                                         UCAR Unidata Program
address@hidden                      http://www.unidata.ucar.edu



Ticket Details
===================
Ticket ID: NFO-217382
Department: Support netCDF
Priority: Normal
Status: Closed