Dplyr: group_indices() cannot be used inside a dplyr expression

Created on 24 Jul 2018 · 4Comments · Source: tidyverse/dplyr

Hi,

I've recently come across the need of labelling groups generated from group_by(...) inside adplyr expression using piping. Issue #2160 led me to issue #1185 and then to the merged feature #3490 dating back from May. Yet this does not seem to be implemented as of v0.7.6. Easily replicated using the code chunk below.

# Not working, throws a 'missing ".data" argument'
df <- mtcars %>% group_by(cyl) %>% mutate(g = group_indices())

# Bypass issue with a non-piped mutate call
df <- mtcars %>% group_by(cyl)
df <- mutate(ungroup(df), g = group_indices(df))

Has this been merged but not yet included in the release ? Or is that unintented ?

Cheers
Thomas

Source

tobadia

👍1

Most helpful comment

I think this is how you actually want to get your result:

mtcars %>% mutate(g = group_indices(., cyl))

Came across this issue because I was having trouble until I figured it out!

benjaminrobinson on 26 Jul 2018

👍3

All 4 comments

I think this is how you actually want to get your result:

mtcars %>% mutate(g = group_indices(., cyl))

Came across this issue because I was having trouble until I figured it out!

benjaminrobinson on 26 Jul 2018

👍3

Indeed that works! However syntax feels contradictory to what is suggested in the merged PR #3490.

tobadia on 26 Jul 2018

The development version of dplyr has not been released to CRAN yet.

hadley on 26 Jul 2018

This old issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with reprex) and link to this issue. https://reprex.tidyverse.org/