You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I really like kerchunk. I started #123 because I wanted to use kerchunk on MPIESM climate model output, which turns out is Netcdf2. kerchunk is built on h5py which I think requires Netcdf4. Are there current efforts to widen the range of input types for kerchunk, i.e. allowing more file types, maybe even these that can be opened by xarray.open_dataset?
Or would this essential mean replicating much of the xarray io again in kerchunk?
The text was updated successfully, but these errors were encountered:
netCDF4/HDF5 was indeed the first format we have developed (largely thanks fo @ajelenak 's deep knowledge of h5py) and certainly the most fruitful. We fully intend to support more formats, and netCDF-classic (i.e., versions 1-3) is definitely on that list.
( @rsignell-usgs , I don't remember which of your suggested datasets included both netCDF4 and netCDF-classic, please link here)
I really like
kerchunk
. I started #123 because I wanted to usekerchunk
on MPIESM climate model output, which turns out isNetcdf2
.kerchunk
is built onh5py
which I think requiresNetcdf4
. Are there current efforts to widen the range of input types forkerchunk
, i.e. allowing more file types, maybe even these that can be opened byxarray.open_dataset
?Or would this essential mean replicating much of the xarray io again in kerchunk?
The text was updated successfully, but these errors were encountered: