Distinct .keep_all parameter

[rrremedio]

Would not it be more natural to the parameter ".keep_all" in distinct function to be value as TRUE? I mean, as the 'd' of dplyr denotes, i'd like to subset the data.frame (here comes the 'd') based on a set of unique values of certain variables or in all of them and keep all columns, that is, keep the data.frame.

[const-ae]

Actually as far as I understand it was the default for distinct to keep all variables until dplyr version 0.5 (release notes). This issue contains more information why they changed the behavior.

[krlmlr]

It's probably not worth the effort to change that, as it's easy to change the behavior or write your own wrapper. @hadley: Please comment.

[hadley]

I don't think it is more natural because the values of the other columns are essentially arbitrary (because row order generally shouldn't be considered meaningful)