Alexander Barth

See also: https://github.com/Unidata/netcdf-c/issues/2049 https://github.com/Unidata/netcdf-c/issues/2402 Some additional context: https://blog.r-project.org/2020/05/02/utf-8-support-on-windows/

I just tried the master version of your NetCDF.jl package and the DiskArray integrations works very nice and natural! Do I understand it correctly, that the disk access will be...

I made a try in a branch `DiskArrays` but I am hitting this error: ``` Got exception outside of a @test ArgumentError: Indices of type Tuple{StepRange{Int64,Int64},StepRange{Int64,Int64}} are not yet Stacktrace:...

> Having issues collected here in this thread is fine. Regarding the Julia compatibility, I just created a branch that checks for the Julia version and for Julia 1.0 only...

> Regarding the strided reading I am not sure what to do. The simplest solution would be to add a method that simply reads to whole array and then returns...

Maybe multiple dispatch can help here? The NetCDF/HDF/Zarr package defines what type of indices it supports: ```julia readblock!(v::ZarrVar, aout, r::AbstractUnitRange...) readblock!(v::NetCDFVar, aout, r::Colon...) # calls nc_get_var (all the data) readblock!(v::NetCDFVar,...

Great, thanks Fabian! I will test the updated version of DiskArrays. I will also make some benchmarks; so far NCDatasets is about 2x faster than Python and R. I hope...

Thanks you for letting me know! I am still trying the update the branch so that I passes the test suite. I have some problem with NetCDF scalars (arrays with...

@meggart, do you have any option whether NetCDF scalar should a 0D-array or a regular julia scalar? When I index the `PseudoDiskArray` example with a 0D-array, I get the following...

Thanks you for the fix! I confirm that 0D-Arrays work now. The next issue are arrays with unlimited dimensions (in netcdf 3.x you can have one and in netcdf 4.x...