[R-pkg-devel] tibbles are not data frames

Hadley Wickham h.wickham at gmail.com
Tue Sep 26 21:48:03 CEST 2017


On Tue, Sep 26, 2017 at 10:40 AM, Joris Meys <Joris.Meys at ugent.be> wrote:
> On Tue, Sep 26, 2017 at 5:33 PM, Hadley Wickham <h.wickham at gmail.com> wrote:
>>
>> > I for one am happy this discussion pops up, because it's a piece of
>> > information I give to my students as well: convert to a data.frame when
>> > you
>> > start your analysis just to play safe. And this discussion shows why
>> > that is
>> > -for the time being!- a good advice. The moment tibbles become the
>> > default
>> > data format in R, or some R++, or in Julia for all I care, I'll be more
>> > than
>> > happy to burn that drop = FALSE on a stake. But for now we can't ignore
>> > the
>> > differences and the potential for conflicts when you try to use a tibble
>> > instead of a data.frame.
>>
>> I think this is sub-optimal advice because most functions do work fine
>> with tibbles.
>
>
> Most. Not all. Either tibbles work exactly like a data.frame, or they don't.
> If they do, I wouldn't give that advice. But they don't.

They work 95% like a data frame. Seems odd to recommend that you
coerce 100% of the time for a <5% of the time problem.

>> It is only a few packages (largely written some time
>> ago) that don't. And typically, if they don't work with tibbles,
>> you'll get a (usually slightly confusing) error message because some
>> function will get a data frame instead of a vector. So as far I can
>> tell, you only need to as.data.frame() retrospectively, not
>> prospectively. Are you aware of any code that returns an incorrect
>> result (i.e. no error) when given a tibble instead of a data frame?
>
>
> x <- tibble(a = 1:5, b = 5:1)
>
> relcount <- function(x, id){
>   table(x[,id]) / length(x[,id])
> }
> relcount(x, "a")
> relcount(as.data.frame(x), "a")
>
> You're welcome.

Obviously you can contrive an example that fails (why wouldn't you use
nrow() here?). I meant an existing function in a package.

Hadley


-- 
http://hadley.nz



More information about the R-package-devel mailing list