Anshul Singhvi

Results 1170 comments of Anshul Singhvi

This particular change breaks one of the CI examples, specifically: ```julia rcopy(R"data.frame(a=1,b=2)") ``` which internally receives a Float64 for each column rather than a vector for each column. I'm not...

Huh! OK, I didn't realize that other classes are able to override `data.table` even if it is first in the vector of classes. For my particular application, I can just...

It would have to be processed by Literate first, no? Otherwise everything will be a header.

Huh! ```julia ras = Raster(AG.getoverview(b, 1), PS.agg_axis.(dims(r), 2^(2))) heatmap(read(ras)) ``` ![display](https://github.com/user-attachments/assets/0be4a9a3-7cb6-4840-89d2-4d799199dc57) ```julia heatmap(ras) ``` ![display](https://github.com/user-attachments/assets/0a04e115-16ff-4dc6-acfd-a0f2ba710a00) This looks like good old DiskArrays collect again.

See kadyb/vector-benchmark#12

by the way, I just tested `proj_trans_generic` on an array (since I've been doing a bunch of reinterpret stuff :D) and it turns out, it is almost exactly the same...

I can't reprodce this, should we close as stale?

`heatmap(rgba_raster)` just works now: ![download-8](https://github.com/user-attachments/assets/29f5874e-e6c3-4c62-9e24-5565e3316d8c)

Can this be closed now that we have `zonal`?

https://github.com/JuliaGPU/GPUArrays.jl/tree/master/lib/JLArrays may be helpful, it sticks to the GPUArrays interface but runs on CPU. Otherwise you could test on macos (which is guaranteed to have a GPU) but Metal.jl is...