[BUG?] filter() throws error after group_by() %>% do()

[antoine-lizee]

It took me a while to create a reproducible example - I hope this is not double posting (see #278 - similar problem supposed to be solved?). It seems that there is a problem with the lazy evaluation:

# Example filtering that works
res <- diamonds %>% filter(carat >= min(diamonds$carat))
# Let's throw in a random grouped operation:
interm <- diamonds %>% group_by(clarity) %>% do(data.frame(.))
# This breaks:
interm %>% filter(carat >= min(diamonds$carat))
# This doesn't:
m <- min(diamonds$carat)
interm %>% filter(carat >= m)

[realAkhmed]

I can confirm that I experience the same issue. Adding ungroup() before filter() solves the issue

interm %>% ungroup() %>% filter(carat >= min(diamonds$carat))

[hadley]

Minimal reprex:

library(ggplot2)
library(dplyr)

interm <- diamonds %>% group_by(clarity) %>% do(data.frame(.))
interm %>% filter(carat >= min(diamonds$carat))
interm %>% ungroup() %>% filter(carat >= min(diamonds$carat))

probably something to do with rowwise

[romainfrancois]

This was a problem in hybrid evaluation.

[antoine-lizee]

Thanks!