Attributes of non key-value columns in spread

[stephematician]

Is there a reason that column attributes are voided when using spread()? e.g.

D <- data.frame(x=c(1,2), y=c('A','B'), z=c(6,7))
attributes(D$x) <- list(a='b')
sapply(spread(D, y, z), attributes)

Gives:

$x
NULL

$A
NULL

$B
NULL

[batpigandme]

Hi,

So, I believe this is related to the handling of attributes in dplyr as well. On that front, there are several issues open that might be worth looking at:
tidyverse/dplyr#3259
tidyverse/dplyr#2752

[hadley]

Because attributes aren't preserved when subsetting:

df <- data.frame(x=c(1,2), y=c('A','B'), z=c(6,7))
attributes(df$x) <- list(a='b')
attributes(df[1:2, ]$x)
#> NULL

^{Created on 2019-01-04 by the reprex package (v0.2.1.9000)}