Data.table and group named "V1"

[matthieugomez]

There is an issue in filter if the dataset object is a data.table and the variable by which the variable is grouped is named "V1".

library(dplyr)
library(hflights)
library(data.table)
DT <- as.data.table(hflights)
DT %>% group_by(DayOfWeek) %>% filter(ArrTime==1502) %>% nrow
[1] 173
setnames(DT,"DayOfWeek","V1")
 DT %>% group_by(V1) %>% filter(ArrTime==1502) %>% nrow
[1] 3239

It happens because of this line of code:

call <- substitute(data[, .I[expr], by = vars])
indices <- eval(call, env)$V1