sl-solution

I guess I can add a little to this: I think the skipmissing(), mean() ... are supposed to be general functions to work with many different data structures in many...

just to add an example to demonstrate my point let me go through a simple example. In the following code there is nothing wrong about the count() function, however, in...

Maybe some of these should be customised for DataFrames, e.g. if we are dealing with one dimensional numeric columns then there are less general but faster way to calculate std...

Just a side question, is there any situation that some one working with DataFrame wouldn't like to deal with missings? (i.e. a scenario that f(x::Vector) return missing when any but...

> I am not sure if I understand the question. Do you ask if there are realistic scenarios in which one would want `missing` to propagate like e.g. in: >...

> I guess this is a commonly agreed standard how functions in statistics realm should work by default. Here is an example from R session: > > ``` > >...

I understand your points and I am ok with them, but I was thinking about something mild like fast path of aggregation in grouped data frame rather than dealing (or...

> I don't understand what this means. Sorry for confusion, let me elaborate this. In `DataFrames` when an aggregation(like sum, mean, ...) is done for a GroupedDataFrame a customised approach...

The other way should be to write a new newton function from the scratch. I think the capability of finding zeros of functions is needed for Distributions.jl (e.g. for obtaining...

I think there are not that many questions to have a separate channel beside [Julia discourse](https://discourse.julialang.org/). Do you see any problem with discourse? And I am not sure how much...