Mutate doesn't work with functions with format pkg::function

[wch]

For example, with dplyr not attached:

dplyr::mutate(mtcars, cyl2 = dplyr::lag(cyl))
# Error in as.character(substitute(name)) : 
#   cannot coerce type 'closure' to vector of type 'character'

dplyr::mutate(mtcars, cyl2 = lag(cyl))
# OK

[smbache]

On the less serious side, one can do

mtcars %>% 
   (c(quote(dplyr::mutate), 
    cyl2 = substitute(f(cyl, 2), list(f = dplyr::lag))) %>% 
    as.call)

[hadley]

@romainfrancois This is another hybrid evaluation buglet, I think. Probably need to hybrid implementation of ::.

[andrewblim]

Using R 3.1, dplyr 0.2, OS X 10.9.4: @wch's second call without the dplyr scoping does not work correctly for me, as the cyl2 column is the same as cyl:

> dplyr::mutate(mtcars, cyl2 = lag(cyl)) %>% dplyr::select(cyl, cyl2)
   cyl cyl2
1    6    6
2    6    6
3    4    4
4    6    6
5    8    8
...

I guess this is because it uses stats::lag and this doesn't work correctly? (is this a bug maybe, is this supposed to work?) However it seems I can force the use of dplyr::lag and get a correct result by surrounding the call in parens, a little easier than @smbache's solution:

> dplyr::mutate(mtcars, cyl2 = (dplyr::lag(cyl))) %>% dplyr::select(cyl, cyl2)
   cyl cyl2
1    6   NA
2    6    6
3    4    6
4    6    4
5    8    6
...

[renkun-ken]

stats::lag() works perfectly with pipeR:

library(pipeR)
mtcars %>>%
  dplyr::mutate(cyl2 = lag(cyl)) %>>%
  dplyr::select(cyl, cyl2) %>>%
  head

  cyl cyl2
1   6   NA
2   6    6
3   4    6
4   6    4
5   8    6
6   6    8

Does not work with dplyr::lag():

mtcars %>>%
  dplyr::mutate(cyl2 = dplyr::lag(cyl)) %>>%
  dplyr::select(cyl, cyl2) %>>%
  head

Error in as.character(substitute(name)) : 
  cannot coerce type 'closure' to vector of type 'character'

Does not find dplyr::lag in ():

mtcars %>>%
  dplyr::mutate(cyl2 = (dplyr::lag(cyl))) %>>%
  dplyr::select(cyl, cyl2) %>>%
  head

Error: 'lag' is not an exported object from 'namespace:dplyr'

[wch]

@renkun-ken The problem is with calling stats::lag() (instead of lag()) in mutate(). (The dplyr::lag() function was recently removed in favor of stats::lag() but the problem remains.)

[renkun-ken]

Thanks @wch for your remind :) NSE problems always look frustrating...