Matthieu Gomez

Could we switch to & and && in the next version (deprecating *), and then switch to * and **.

Bump. I would really support using one character (any character!) to denote interaction and two characters to denote main effects and interactions. Also, I think && is really fine. The...

Environment with the best performances ``` (Try) pkg> status Status `~/FormerDataFrames/Project.toml` [a93c6f00] DataFrames v0.17.1 ⚲ [3eaba693] StatsModels v0.5.0 ⚲ ``` Contrary to what I said in my previous posts, I...

I'm on the same computer anymore. On the most recent version of StatsModels v0.6.11 DataFrames v0.20.2 ```julia julia> @benchmark ModelMatrix(ModelFrame($f, $df)) BenchmarkTools.Trial: memory estimate: 764.15 MiB allocs estimate: 226 --------------...

It could also use automatic sparsity detection right?

One syntax could be something like ```julia function OnceDifferentiable(f!, x0, F0, J0; color = matrix_colors(J0), autodiff = true) if autodiff == true j! = (J, x) -> DiffEqDiffTools.finite_difference_jacobian!(J, f!, x,...

It allows the user to indicate the sparsity structure of the Jacobian (which is then used by `matrix_colors` and the functions that create the Jacobian in place). That being said,...

To complete my previous post, R misses a set of functions that allow to compute efficiently - moments mean, sd, var, cov, skewness, kurtosis - quantile type 3: min, max...

Thanks for the answer, it's very interesting. About your last paragraph. Even with raw vectors and without weight/selection, it seems like `colRanges` is faster than `range` for integers or doubles....

Yes I think that a differentiate method with three states would be enough. It would be a great addition!