r - How does dplyr::do work with data.table -

the dplyr::do not seem work data.table:

# works data.frame(1) %>%    do(data.frame(1)) ##   x1 ## 1  1 # same data.table not work data.table(1) %>%    do(data.frame(1)) ## error in do_.data.table(.data, .dots = lazyeval::lazy_dots(...)) :  ##  argument ".f" missing, no default 

some investigation lead functions do , do_.data.table:

do ## function (.data, ...)  ## { ## do_(.data, .dots = lazyeval::lazy_dots(...)) ## } ## <environment: namespace:dplyr> dplyr:::do_.data.table ## function (.data, .f, ...)  ## { ## list(.f(as.data.frame(.data), ...)) ## } ## <environment: namespace:dplyr> 

how work? arguments of do_.data.table not compatible gets do. , result of do_.data.table list instead of data.frame. how use do or do_ data.table input?

i know can use df %>% data.frame %>% do(...), hoping direct solution.

it looks do_ .f arguments , return in list.

data.table(1) %>%     do_(data.frame(2), data.frame(3), .f = function(x1, x2, x3) cbind(x1, x2, x3))  # [[1]] #   v1 x2 x3 # 1  1  2  3 

don't know do though.