Due to the current gap in continued funding from the U.S. National Science Foundation (NSF), the NSF Unidata Program Center has temporarily paused most operations. See NSF Unidata Pause in Most Operations for details.
Heiko Klein wrote:
Hi John,thanks for your feedback concerning grib/BUFR. Reading your argument about the netcdf-library being able to read grib-data, I don't think it will be a problem for us when the collaboration will decide for grib, though I still will prefer the netcdf-library.Today, I ran into another problem with reading netcdf-files with 2.2.09 and 2.2.10 . At first I got a EOFException: at ucar.unidata.io.RandomAccessFile.readInt(RandomAccessFile.java:835) at ucar.unidata.io.RandomAccessFile.readFloat(RandomAccessFile.java:945)at ucar.nc2.N3raf.readData(N3raf.java:81) at ucar.nc2.N3iosp.readData(N3iosp.java:111) at ucar.nc2.NetcdfFile.readData(NetcdfFile.java:965) at ucar.nc2.Variable._read(Variable.java:758) at ucar.nc2.Variable.read(Variable.java:524) at international.emep.io.NetcdfUtils.copy(NetcdfUtils.java:57)at international.emep.io.GDVtoCF1_0converter.<init>(GDVtoCF1_0converter.java:76) at international.emep.io.GDVtoCF1_0converter.main(GDVtoCF1_0converter.java:169)(linenumbers might not be exactly the same as yours, since I included up to 4 println statements.)The error occurs when reading the last variable from the file with var.read():variables = ncIn.getVariables(); for (Variable var : variables) { try { System.err.println("Writing variable: "+var.getName()); ncOut.write(var.getName(), var.read()); System.err.println("finished"); } catch (InvalidRangeException ire) {throw new IOException("InvalidRangeException: "+ire.toString());} }Looking at the copied file with ncview, I see that all data looks misplaced. With the println I got for the last variable:Writing variable: WDEP_RDN Trying to read 14652 floats Fetching 58608 starting at 4138964, but filesize is just: 4194340
Can you send me the file or a URL to it? How was the file written? From the java library or the c library?
BTW: I've seen that unidata has a netcdf mailing list. Do you prefer me to send such messages to the list, or directly to you?
Ive cc'ed it to netcdf-java, which is a good place to send all such problems, so that others can "follow along" if they wish.
Best regards, Heiko
thanks for your help debugging.
netcdf-java
archives: