Processing nested lists in R - r

Question

I have a nested list with the below structure:
Each $mu is a vector of 15 numbers. How can I calculate the average of each of the 15 elements across 1000 $mu's?
I know I can access the $mu's with hb.post.PT1$compdraw[[1]]$mu, but how can I extract or loop over them to execute functions?
I've been trying stuff like this without success:
lapply(hb.post.PT1$compdraw, function(x) lapply(hb.post.PT1$compdraw[[x]]$mu, function(x) mean))
My apologies, I cannot post example list data.

Answer 1

You are already looping through each list element with the first lapply. So based on your structure each iteration of the loop will be a list of 2 elements (mu and rooti). So you only need one lapply:
ll <- list(list(mu = runif(15), root = runif(10)), list(mu = runif(15), runif(10)))
str(ll)
#List of 2
# $ :List of 2
# ..$ mu : num [1:15] 0.314 0.614 0.228 0.594 0.447 ...
# ..$ root: num [1:10] 0.73 0.496 0.266 0.599 0.917 ...
# $ :List of 2
# ..$ mu: num [1:15] 0.134 0.244 0.507 0.771 0.934 ...
# ..$ : num [1:10] 0.818 0.138 0.83 0.509 0.914 ...
lapply(ll, function(x) mean(x$mu))
#[[1]]
#[1] 0.4264001
#
#[[2]]
#[1] 0.4526724

Answer 2

You can use
hb.post.PT1$compdraw %>%
purrr::map(function(x) {
mean(x$mu)
}) %>% unlist

Answer 3

This is how I would do it with purrr::imap from the tidyverse. We can apply a function that creates a tbl with the list index and the mean of that index's mu vector, and bind them into one dataframe for easy display.
library(tidyverse)
set.seed(12345)
testlist <- list(
list(mu = rnorm(15), rooti = rnorm(15)),
list(mu = rnorm(15), rooti = rnorm(15)),
list(mu = rnorm(15), rooti = rnorm(15)),
list(mu = rnorm(15), rooti = rnorm(15)),
list(mu = rnorm(15), rooti = rnorm(15))
)
str(testlist[1:2])
#> List of 2
#> $ :List of 2
#> ..$ mu : num [1:15] 0.586 0.709 -0.109 -0.453 0.606 ...
#> ..$ rooti: num [1:15] 0.817 -0.886 -0.332 1.121 0.299 ...
#> $ :List of 2
#> ..$ mu : num [1:15] 0.812 2.197 2.049 1.632 0.254 ...
#> ..$ rooti: num [1:15] 1.461 -1.413 0.567 0.583 -1.307 ...
testlist %>%
imap_dfr(.f = ~ tibble(index = .y, mu_mean = mean(.x$mu)))
#> # A tibble: 5 x 2
#> index mu_mean
#> <int> <dbl>
#> 1 1 0.0341
#> 2 2 0.448
#> 3 3 0.146
#> 4 4 -0.130
#> 5 5 0.244
Created on 2018-05-07 by the reprex package (v0.2.0).

Answer 4

Could be overkill for this case but for this kind of problem in general it's good to remember about purrr::transpose
map(transpose(hb.post.PT1$compdraw)$mu,mean)