Andy Dienes
Andy Dienes
any of ``` reshape(a, axes(b)) reshape(a, axes(b)...) reshape(b, size(b)) reshape(b, size(b)...) ``` all work right now and are equal, I think? would feel a little inconsistent if the curried version...
I see. why not add a dispatch for ``` reshape(axes::AbstractUnitRange...) ``` ?
ah I see, sorry for missing that at first hmm I think your option 2. strikes me as best `Change to always require a tuple, reshape((2, :))` that way there's...
https://github.com/JuliaLang/julia/issues/36235
should be fixed now in https://github.com/JuliaArrays/StaticArrays.jl/releases/tag/v1.6.4 ?
I see. The first example, where `predict(mach, df)` takes the first 5 cols, while not ideal, I can see an argument for being an acceptable side-effect of genericity. However in...
I just realized there is the `lag` function. this mitigates the issue of the number of characters somewhat, but the other two annoyances stand maybe `DataFrames` can "re"-export this function?...
I think it's good to stay simple/consistent. that `df.a = df.b` aliases can be a bit surprising, but once you know about it it's fine. I think it would be...
I have not thought out all the edge cases yet, but I will try to get a list of examples for these scenarios (and then bikeshed the actual api) at...
actually, sorry, could you help me create such an edge case that may be ambiguous? maybe this is a naive answer, but I am thinking that in the same way...