Rename rowmean_n()?

[etiennebacher]

From @IndrajeetPatil's comment: 6e3182c#commitcomment-124303628

Sorry, I didn't get to weigh in before, but I personally find the function name very dissatisfying, and even a bit misleading. I think the name should make a reference to the fact that this is relevant for missing values. I know that n is mandatory, and so users may figure it out once they use it mistakenly.

How about renaming it to rowmean_n_missing()?

@strengejacke

[etiennebacher]

On the other hand, if you see rowmean_n() for the first time, what do you think the n stand for?

rowmean_n_missing() is also a bit misleading IMO because n actually refers to the minimal number of non-missing values required to compute the mean.

[etiennebacher]

I guess if I were a user discovering this function:

• rowmean_n() -> hard to know that it does because I don't know what the n stands for, so need to check the docs
• rowmean_n_missing() -> this function computes the rowmean if there are n or more missing values (while it's the opposite)

[IndrajeetPatil]

When I read that name, my first hunch is that it's going to calculate rowwise mean for the first n rows.

[strengejacke]

I don't think there is a perfect name for this function, but rowmean_n_missing() looks very cryptic to me.

[strengejacke]

See also https://r4stats.com/2014/09/03/adding-the-spss-mean-n-function-to-r/

[DominiqueMakowski]

Wouldn't it be easier with just a homemade version of row_mean() (or average_cols()) with some select/exclude sugar and anmax_missing argument?

[strengejacke]

That sounds good!

[strengejacke]

max_missing or min_valid? I'm used to think in minimum amount of valid values...

[etiennebacher]

min_valid sounds better to me