NOTE: The decoders
mailing list is no longer active. The list archives are made available for historical reasons.
On Tue, 3 Jun 2003, Yan Gu wrote:
Robb, I have made a table and use it to convert. There is no problem to make the cdl file. However, it gives me such error messages when I try to make nc file:
Yan, The problem is the udunits package doesn't recognize some of the units for the parameters. I'll note the units to use below for the errors.
---------------------------------------------------------------------------- udunits(3): syntax error: -
replace the "-" with "" in the table
^ udunits(3): syntax error: (0 - 1)
replace "(0 - 1)" with "bit" in the table same for the others. If you have problems with the other variables you can comment out the udunits attribute line with a "//" in the cdl file. Robb...
^
udunits(3): syntax error: (0 - 1) ^ udunits(3): syntax error: - ^ Segmentation fault ----------------------------------------------------------------------------- I have attached the tab file the the cdl file. I also attached the log file which logged the convert infomation. Would you please help me on this error? Yan ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Yan Gu Sr. Computer Program Analyst Center For Ocean-Atmospheric Prediction Studies Florida State University Express Mail: 2035 E. Paul Dirac Drive R. M. Johnson Bldg. -Suite 233 Tallahassee, FL32310 Tel: (850)644-6935 (OFFICE) (850)574-9877 (HOME) ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ On Mon, 2 Jun 2003, Robb Kambic wrote: > On Mon, 2 Jun 2003, Yan Gu wrote: > > > Robb, > > > > Thanks for your answer to me. > > > > For the gribtocdl, now I know that "3008" stands for 3008 parameters. > > However, I am confused that if there are 3008 parameters, then why in the > > result netCDF file, there are only about 47 parameters? Futher more, for > > Yan, > > No there are 47 parameters in the file, some of the parameters have have > different levels, times, etc. Each product only contains one parameter > and level time information and there are multi levels for one parameter. > > > > the table files, now I know the syntax of them. My new question is how to > > contruct the table file. I mean, where I can get the infomation of grib ID > > ( I guess I can get this in gribdump -h ), netCDF name, long name, units > > and comments assume I only have grib files? > > > > I looked up you parameters, they come from a ecmwf parameter table. It > looks like these names are not the standard ones but are close. I'll attach the > table, it's in XML format. > > > Using your ctl file entries, ie > > no10Usfc 0 165,1,0 ** 10 metre u wind component m s**-1 > > transforms into a table entry > > Grib;netCDF ;long name ;units ;comments > 165 ;no10Usfc;10 metre u wind component;m s**-1 ; > > You need to do this for all 47 variables. Then use the table for both > programs gribtocdl and gribtonc. Then the names should be correct in the > NetCDF file. > > The ctl file and parm_table_ecmwf_128.xml helped me decide the numbers. > > Robb... > > > > For the gribtonc, I tried to attach the grib file i am using to > > convert. I am too sad it is too big about 300 MB. I think I had better not > > attach it. Well, I do attach the test.cdl file which from gribtocdl. I > > also attach the Y63211.ctl file which I got from the site as the same time > > as I downloaded the data. Those parameters in that file are what we need > > but for some reason I can't match them in the cdl file that created by > > gribtonc. I guess the problem here is the name difference. But I do have > > no idea how to match them. Would you please give me some clue? > > > > Again, I do appreciate your help. It really help me a lot. > > > > Yan
============================================================================== Robb Kambic Unidata Program Center Software Engineer III Univ. Corp for Atmospheric Research rkambic@xxxxxxxxxxxxxxxx WWW: http://www.unidata.ucar.edu/ ==============================================================================
decoders
archives: