Kai Mühlbauer

@zoj613 Could you try to provide `kwargs=kwargs`? Where kwargs is a dict with method-key inside.

Nevermind, I should read the complete issue, not only the last comment. I think this is just a problem with the doc-string then.

@bnlawrence Yes, thanks, for sure. I've been following the recent activity over at pyfive, great work! A pure python netcdf4/hdf5 backend would be really neat. Looking forward to this.

I'm closing this now. Feel free to re-open with additional detail or open a new issue.

Thanks Ryan for the quick reply. This is in the global attributes (XXXX added by me): ``` Info : Monthly ERA-Interim data. Downloaded and edited by fabien.maussion@XXXXXXX ``` Maybe @fmaussion...

Xarray is able to write these attributes, but for int16 dtype NaN does not make any sense. I'd consider that a bug. But maybe someone with more knowledge on ERA5...

I've created a C-version of the problem. It might gain more attraction here in netcdf-c :grin: ```c #include "hdf5.h" #define H5FILE_NAME "test_dim_scale_null.h5" #define X "x" #define Y "y" #define N1...

> > So, if I understand @itcarroll correctly, the proposal is to allow HDF5 Datasets with H5S_NULL dataspace to be handled as netCDF4-dimensions. > > If you meant to say...

@andreas-schwab Thanks again for this PR. It turned out it was a bit more involved to make this work and we hope #7827 solves the underlying issue.

@felixonmars If you are still in the works with this, I'd appreciate if you could test this against #7827. Thanks.