This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.
Hi Simon- Simon Elliott wrote:
WHen I look at the file attached by HansPeter (HP GRIB.001) it looks fine to me. The GRIB part can be found after the first 899 bytes. The remaining part (GRIB to 7777, inclusive) is 803521 bytes long, which is as indicated in octets 9 to 16 of Section 0.
I'll work with Robb on this next week. I think he made some changes to skip the header, but I don't have access to his changes so it still fails in what I have.
The users who have been looking at the data have received it via EUMETCast, and also via ftp from me. Maybe some got data from the UMARF, but I don't get to hear about them. I think that as long as a decoder looks through a file and pulls out GRIB (or BUFR) messages before decoding, then it doesn't matter how they are packed into files.
I agree in principle. It's fine if you are passing a file to a decoder and are assured that the decoder is a GRIB decoder and that the user says that there is a GRIB message in the file. However, in our case, we have a top level data handler that has to determine which decoder to pass it to. This top level handler can open netCDF files, GRIB1, GRIB2, OPeNDAP remote URLs, etc. So, it has to make some determination at the start about what is in the particular file. Usually, it does this by looking at the first few bytes of a file. For a netCDF file, it looks for the netCDF header, for a GRIB file, it looks for GRIB1 or GRIB2. The question is how far into the file do you look before giving up? Also, should a decoder have to handle both GRIB and BUFR in the same file? Anyway, we'll wait for Robb's return and work on this more at that time. Thanks for your help and input. Don ************************************************************* Don Murray UCAR Unidata Program address@hidden P.O. Box 3000 (303) 497-8628 Boulder, CO 80307 http://www.unidata.ucar.edu/staff/donm *************************************************************