-
ok on @rabernat encouragement I'm just going to ask rather than do my homework for some burning questions: I have 14000 simple netcdf files that are a regular grid with typical 1D lon,lat arrays and a singleton time slice that is predictable with a file per time.
I think I saw this mentioned somewhere, I'm super new to python but not to programming or netcdf 🙏 |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 2 replies
-
Thanks for asking this question Michael. The feature you want is not directly supported by Xarray per se. However, it is possible via the Kerchunk standard and associated python package.
Basically, you create a virtual Zarr dataset that maps each data chunk from the underlying files to a Zarr chunk name. You can then open that in Xarray. Here's a recent blog post that demonstrates what this enables: https://medium.com/pangeo/accessing-netcdf-and-grib-file-collections-as-cloud-native-virtual-datasets-using-kerchunk-625a2d0a9191 |
Beta Was this translation helpful? Give feedback.
Thanks for asking this question Michael.
The feature you want is not directly supported by Xarray per se. However, it is possible via the Kerchunk standard and associated python package.