Could select() substitute mutate() to add new variable?

[yanlinlin82]

I guess this habit should come from the SQL thinking.

When I occasionally tried the following code:

data.frame(A = 1:3, B = 2:4) %.% select(C = A + B)
in which case the code should be:
data.frame(A = 1:3, B = 2:4) %.% mutate(C = A + B) %.% select(C)
R environment crashed immediately.

I wonder if there will be an improvement to enhance select() to substitute mutate(), or at least providing a warning message for this case rather than crash.

[romainfrancois]

Maybe this can be handled on the R side.

> data <- data.frame(A = 1:3, B = 2:4)
> select_vars( names(data), C = A+B )
 C
NA

I can at least make select_impl handle NA by throwing an error.

[hadley]

Duplicate of #302