Expand data frame for all combinations of values for all variables without explicitly specifying variables names

[mattwarkentin]

Hi,

I posted the following question on Twitter:

I can’t seem to figure out how to use expand to get all combinations of values for all variables in df without specifying each var name explicitly? It’s easy with expand.grid. tibble(x=1:3, y=1:3) %>% expand.grid(). Can I tell expand() to use all vars?

@hadley asked if I would file an issue so he could think about it some more.

I suppose the Tweet is self-explanatory, I want to be able to expand a data frame for all possible combinations of values for every variable, without having to specify variable names explicitly.

This works:

tibble(x = 1:3, y = 1:3) %>% expand.grid()

This doesn't:

tibble(x = 1:3, y = 1:3) %>% expand()

Of course, in the two variable setting, explicitly typing tibble(x = 1:3, y = 1:3) %>% expand(x, y) is not a problem, but it doesn't scale well.

Something like this might be a possible solution:

tibble(x = 1:3, y = 1:3) %>% expand(vars(everything()))

[hadley]

You can use tibble(x = 1:3, y = 1:3) %>% expand_grid() in the dev version.

In the future, we'll support df %>% expand(something(everything()) but that'll come to dplyr first (and requires a little thought).