Cameron Bieganek

FWIW, I've noticed in the past that `dev`-ing a relative path doesn't work very well, independently of VS Code. It's been a while since I dealt with this, but if...

[This proposal](https://github.com/JuliaAI/MLJTuning.jl/issues/199) for MLJTuning would allow MLJ to handle immutable pipelines.

> While the Polars result is technically correct To be clear, the Polars result is *incorrect*. The expected result set is this: ```python (A, 10, 100) (B, 20, 90) (A,...

The order of the output tuples is irrelevant, since your SQL query does not include a global `ORDER BY` clause. What is important is the relationship between the `SalesCumulative` values...

My opinion is that Polars need not sort the output dataframe if there is no global `ORDER BY` clause in the SQL expression. Ordered output should only be guaranteed when...

If we set aside learning networks, it seems like we don't really need `evaluate!`. The non-mutating `evaluate` is sufficient and more intuitive. Is there a way to write a non-mutating...

Yes, the other columns would stay as is. I'm not actually requesting the output dataframe to be sorted in any way. So, the ideal output would be this: ```python shape:...

> Which looks correct to me. Columns `g`, `t`, and `x` remain in their initial ordering, and `x_lag` is computed as if the dataframe were first sorted on `g` and...

> The calculation worked just fine. No. The correct result set (which you can get by *first* sorting the whole data frame, and *then* applying the window function) for `g==2`...

> Right, but we're not talking about a result set here If we're not allowed to talk about result sets, then it becomes quite difficult to talk about table operations...