filter_ with a data.table throws 'argument ".dots" is missing, with no default' error

[richierocks]

To reproduce:

library(data.table)
library(dplyr)
DT <- data.table(x = rnorm(10))
DT %>% filter_(~ x > 0)
## Error in lazyeval::common_env(.dots) : 
##   argument ".dots" is missing, with no default

The error persists if you replace the formula with a string or call:

DT %>% filter_("x > 0")       # throws same error
DT %>% filter_(quote(x > 0))  # throws same error

...and if you don't use piping:

filter_(DT, ~ x > 0)          # throws same error

The interactive version is OK:

DT %>% filter(x > 0)          # OK

[jakesherman]

I am having the same issue. filter_ wasn't working with data.table, but works as intended for data.frame

[paljenczy]

Confirmed, same behavior here. Using dplyr 0.4.1 and data.table 1.9.4.

[richierocks]

Just had another look at this, and it's an easy fix. In filter_.data.table, change the second line

env <- lazyeval::common_env(.dots)

to

 env <- lazyeval::common_env(dots)