Ability to specify lists, Not lists, colon, or nothing for @pivot_longer

[cnrrobertson]

This will allow things like:

@pivot_longer(df, [col1, col5])

or

@pivot_longer(df, -[col1, col5])

and my favorites:

@pivot_longer(df, :)

and

@pivot_longer(df) # Equivalent to :

This is not consistent with the tidyr syntax which would be pivot_longer(df, cols=everything()) but I think it's more ergonomic and Julian. Let me know if you think it should be more tidyverse-esque.

Addresses #10.

[kdpsingh]

Appreciate your work on this @cnrrobertson! Have a tight deadline on another project so will look at this in 2 weeks.

In general, my philosophy is that the syntax should look like tidyverse. We include an everything() function in TidierData which is merely an alias for All(). I'm curious to see how the multi-negated selection is implemented. Will take a look and share feedback.

[cnrrobertson]

Aha! I had no idea about everything(). I actually authored this PR just to try to solve that very issue. Should have asked on Slack first..

Either way, please look it over whenever you have time. I'm fairly new to macro programming, but these were actually very straightforward implementations. Basically just grabbing vectors or negated vectors in @capture.

I should mention that this is currently limited and won't work for things like: @pivot_longer(df, [col1, col10:col12]).

I have also worked on an implementation that would allow splatted names like @pivot_longer(df, col1, col3, col5) but it gets tricky with arbitrary combinations like @pivot_longer(df, col1, col7:col8). Let me know if this is a direction you would like to pursue.

[kdpsingh]

Thank you again for working on this. This is great! There are a few edge cases I need to fix with the negated selection and then I will merge.

[kdpsingh]

I should mention that this is currently limited and won't work for things like: @pivot_longer(df, [col1, col10:col12]).

It does now!