yt87
commented
3 weeks ago The code on GitHup is old, not compatible with current xarray. I stopped working on it. In retrospect, the main idea was wrong. The GRIB2 and NetCDF (which is what xarray uses) models are not compatible. In NetCDF, a variable is identified by a name, with attributes. In GRIB2, iit is defined by Section3: https://www.nco.ncep.noaa.gov/pmb/docs/grib2/grib2_doc/. For example:TMP is temperature. It can be at a given level, within a layer, average in a time period and so on. All those are different variables and have to be defined in the NetCDF model with different names. Pywgrib2 attempts to extend the name, for example TMP_100mb. However, this approach is unwieldy, there are simply too many tags to be added.
I recommend cfgrib, it is well integrated with xarray. If you use GRIB2 files made by NCEP, you'll have to split them before passing them to cfgrib. At least cfgrib will tell you when there is name conflict, instead of attempting ugly variable names. Which does not always work anyway.
Le lun. 7 oct. 2024, à 17 h 45, stackjohn @.***> a écrit :
Hello, is there a guide which explains how to build this without conda?
I get this:
clang: error: no such file or directory: 'pywgrib2_xr/_wgrib2.c'I'm not sure where the source of that file should be.
— Reply to this email directly, view it on GitHub https://github.com/yt87/pywgrib2_xr/issues/5, or unsubscribe https://github.com/notifications/unsubscribe-auth/AJS3CCZEXAPC5ZVXGWGSITTZ2LCEBAVCNFSM6AAAAABPQOQMAGVHI2DSMVQWIX3LMV43ASLTON2WKOZSGU3TCMBXGI4DANA . You are receiving this because you are subscribed to this thread.Message ID: @.***>
Hello, is there a guide which explains how to build this without conda?
I get this:
I'm not sure where the source of that file should be.