tidyverse - delete a column within a nested column/list

问题

I have the following data:

(Note: I'm using the current github version of dplyr within tidyverse which offerse some new experimental functions, like condense - which I'm using below, but I think that's not relevant for my problem/question).

library(tidyverse)
library(corrr)

dat <- data.frame(grp = rep(1:4, each = 25),
                  Q1 = sample(c(1:5, NA), 100, replace = TRUE),
                  Q2 = sample(c(1:5, NA), 100, replace = TRUE),
                  Q3 = sample(c(1:5, NA), 100, replace = TRUE),
                  Q4 = sample(c(1:5, NA), 100, replace = TRUE),
                  Q5 = sample(c(1:5, NA), 100, replace = TRUE),
                  Q6 = sample(c(1:5, NA), 100, replace = TRUE))

I now want to calculate the correlation between Q1 to Q6 within each group, and I'm using:

cor_dat <- dat %>%
  group_by(grp) %>%
  condense(cor = correlate(cur_data()))

Which gives me the correlations as a list-column (?). Within each list, the first column is called rowname and I want to simply delete this column from each list in a tidyverse way. How can I do this?

I already tried something naive like select (-rowname), but this doesn't work.

回答1:

The suggestion by @r2evans would work if we remove the group attribute

library(dplyr)
library(purrr)
cor_dat %>%
     ungroup %>%
     mutate(cor = map(cor, ~ select(.x, -rowname)))
# A tibble: 4 x 2
#    grp cor             
#  <int> <list>          
#1     1 <tibble [6 × 6]>
#2     2 <tibble [6 × 6]>
#3     3 <tibble [6 × 6]>
#4     4 <tibble [6 × 6]>

When there is a group attribute, it results in error

cor_dat %>% 
    mutate(cor = map(cor, ~ select(.x, -rowname)))

Error: mutate() argument cor errored. ℹ cor is map(cor, ~select(.x, -rowname)). ℹ The error occured in row 1. ✖ no applicable method for 'select_' applied to an object of class "character" Run rlang::last_error() to see where the error occurred.

which is consistent with the same behavior if we extract as a column

cor_dat$cor %>% 
           map(~ .x %>% select(-rowname))

Or if we want to make it shorter, it can be done within condense itself because correlate adds a rowname column as per the documentation

dat %>%
  group_by(grp) %>%
  condense(cor = correlate(cur_data()) %>%
                              select(-rowname))
# A tibble: 4 x 2
# Rowwise:  grp
#    grp cor             
#  <int> <list>          
#1     1 <tibble [6 × 6]>
#2     2 <tibble [6 × 6]>
#3     3 <tibble [6 × 6]>
#4     4 <tibble [6 × 6]>

回答2:

Oh my, it was as simple as:

cor_dat <- dat %>%
  group_by(grp) %>%
  condense(cor = correlate(cur_data())) %>%
  mutate(cor = list(cor[,-1]))

Got there by trial & error. Although this only works with column positions, not names.

来源：https://stackoverflow.com/questions/60270144/tidyverse-delete-a-column-within-a-nested-column-list

标签

nested

tidyverse