Applying purrr::map over each of a vector of characters

问题

I'm working on trying to understand the purrr::map function a little better. Let's say I have a simple vector of characters, and I want to run some function that outputs a data frame using each character as an input.

Here's a toy example

animals <- c('sheep', 'cow', 'horse')

make_df <- function(x){
    data.frame(r1 = rnorm(1:5), r2 = rnorm(1:5), an = x)
}

Here's what make_df

> make_df('sheep')
r1          r2           an
-0.18069698 -0.4767575  sheep
0.09580225  0.2785548   sheep
-0.74701529 0.2673391   sheep
-1.62795239 1.0026010   sheep
0.36573951  -0.2323944  sheep

Now, I want to run this function for each animal and save each dataframe to a list, and put that list in a new data frame where one column is the animals and one column is the list of data frames (which I'll use with other tidyverse functions).

I think the way to do that would be something like data.frame(animals = animals) %>% mutate(ldf = map(animals, make_df(.)))

 data.frame(animals = animals) %>% mutate(ldf = map(animals, make_df(.)))

but this gives me an error

Error in mutate_impl(.data, dots): Evaluation error: arguments imply differing number of rows: 5, 3. Traceback:

1. data.frame(animals = animals) %>% mutate(ldf = map(animals, make_df(.)))
2. withVisible(eval(quote(`_fseq`(`_lhs`)), env, env))
3. eval(quote(`_fseq`(`_lhs`)), env, env)
4. eval(quote(`_fseq`(`_lhs`)), env, env)
5. `_fseq`(`_lhs`)
6. freduce(value, `_function_list`)
7. withVisible(function_list[[k]](value))
8. function_list[[k]](value)
9. mutate(., ldf = map(animals, make_df(.)))
10. mutate.data.frame(., ldf = map(animals, make_df(.)))
11. as.data.frame(mutate(tbl_df(.data), ...))
12. mutate(tbl_df(.data), ...)
13. mutate.tbl_df(tbl_df(.data), ...)
14. mutate_impl(.data, dots)

Of course, I can make this list of data frames with lapply

dfs <- lapply(animals, make_df)

But then if I try to bind dfs and animals into one data frame, I get a seemingly related error, again about differing numbers of rows.

 data.frame(animals, dfs)

Error in data.frame(animals, dfs): arguments imply differing number of rows: 3, 5 Traceback:

1. data.frame(animals, dfs)
2. stop(gettextf("arguments imply differing number of rows: %s",   .     paste(unique(nrows), collapse = ", ")), domain = NA)

Clearly, I am missing something fundamental here. Why am I unable to combine a list of characters and a list of data frames into one data frame manually, and how am I misusing purrr::map so that R runs into a problem, presumably trying to do the same within my functions.

回答1:

Your syntax is a little bit off, you'd either use map(animals, make_df) or map(animals, ~ make_df(.)), the second argument of map needs to be a function, which is the same as lapply:

data.frame(animals) %>% mutate(ldf = map(animals, make_df)) %>% as.tibble()
# A tibble: 3 x 2
#  animals                  ldf
#   <fctr>               <list>
#1   sheep <data.frame [5 x 3]>
#2     cow <data.frame [5 x 3]>
#3   horse <data.frame [5 x 3]>

data.frame(animals) %>% mutate(ldf = map(animals, ~ make_df(.))) %>% as.tibble()
# A tibble: 3 x 2
#  animals                  ldf
#   <fctr>               <list>
#1   sheep <data.frame [5 x 3]>
#2     cow <data.frame [5 x 3]>
#3   horse <data.frame [5 x 3]>

Or if using the data.frame constructor, you need to use I to create a list type column:

data.frame(animals, ldf = I(lapply(animals, make_df)))
#                         ^

来源：https://stackoverflow.com/questions/48144262/applying-purrrmap-over-each-of-a-vector-of-characters

标签

list

dataframe

purrr