cbind'ing grouped df and df gives a matrix?

[hongooi73]

This is... odd behaviour to say the least.

grp_df <- data_frame(x1 = 1:10, x2 = 1:10) %>% group_by(x1)
y_df <- data_frame(y = 1:10)

foo <- cbind(grp_df, y_df)

foo
#   grp_df               
#x1 Integer,10 Integer,10
#x2 Integer,10 Integer,10

foo[[1, 1]]
# [1] 1 2 3 4 5 6 7 8 9 10

This is with dplyr 0.5.0.9000, built on 24-Jan-2017.

[hadley]

Duplicate of #2138. As far as I can tell, dplyr already provides the correct methods, but rbind()/cbind() don't call them because it uses a special dispatch mechanism that doesn't work for data frame subclasses.

[hongooi73]

I just rebuilt dplyr, and the problem is still there. Is there any solution besides just calling cbind.data.frame directly?

[hadley]

Yes, call bind_cols()