Alexander Barth

Indeed, it is not easy to implement this efficiently. Maybe one can implement two sets of types (for NetCDF variable (NcVarRaw, NcVarScaled) and NetCDF files (NcFileRaw, NcFileScaled). Indexing a NcFileRaw...

In case, you are interested. Here is my attempt: https://github.com/Alexander-Barth/NCDatasets.jl

In fact, I started to change netcdf_c.jl quite bit and I am not sure if you are interested in these changes. I think that this file is an automatically generated...

Ok, I agree. The other aspect (similar to scaling) that I have implemented is returning time variables (if they have a unit attribute of the form "days/hours/.. since start date)...

Initially, I considered indeed to depend NCDatasets.jl on NetCDF.jl but I did not choose to follow to road: * the interface in netcdf_c.jl is a bit inconveniet to use e.g....

As a test, I have enabled zstandard in these these builds, if you want to try (it is not yet available in Yggdrasil): https://github.com/Alexander-Barth/NetCDF_jll.jl/releases/tag/NetCDF-v400.902.19%2B0 NetCDF is now build with these...

> If the _FillValue attribute is not set in a NetCDF file, it seems this package defaults to all arrays being read with Union{Missing,T} eltypes, i.e. a missing fill value....

OK, but in this example `_FillValue` is indeed set. We use `missing` because, NaN does not work for integers for example. In your case, what you can do is one...

> fillvalue = nothing ignores the fillvalue that is set, right? It is correct. If somebody would want to extend `cfvariable` to support e.g. ``` cfvariable(ds.group["posterior"],"g",sentinelvalue = NaN) ``` where...

Last time I tried, DiskArrays in NCDatasets I had some issues with types of undefined size like String and probable also NetCDF4 variable-length arrays: ``` julia> sizeof(String) ERROR: Type String...