CODEWITHSUNDEEP
rpurrr
ferrelwill
ferrelwill

Reputation: 821

Different results with map function

I am getting different results with the map function from the library purr. Here is the following code used in Hadley's book, R for data science. Why is this?

code1

models1 <- mtcars %>% split(.$cyl) %>% 
  map(function(df) lm(mpg ~ wt, data=mtcars))

models1 %>% 
  map(summary) %>% 
  map_dbl(~.$r.squared)

code1- result

        4         6         8 
0.5086326 0.4645102 0.4229655

code2

models2 <- mtcars %>% split(.$cyl) %>% 
  map(~lm(mpg ~ wt, data=.))

models2 %>% 
  map(summary) %>% 
  map_dbl(~.$r.squared)

code2-result

        4         6         8 
0.7528328 0.7528328 0.7528328

Upvotes: 0

Views: 64

Answers (1)

dylanjm
dylanjm

Reputation: 2101

The problem is that you are passing the entire dataset mtcars to the lm() function in code-1 instead of passing in df which you declare in your function definition in map(function(df) ...) if you change mtcars to df your problem is fixed

library(tidyverse)

models1 <- mtcars %>% split(.$cyl) %>% 
  map(function(df) lm(mpg ~ wt, data=df))

models1 %>% 
  map(summary) %>% 
  map_dbl(~.$r.squared)
#>         4         6         8 
#> 0.5086326 0.4645102 0.4229655

models2 <- mtcars %>% split(.$cyl) %>% 
  map(~lm(mpg ~ wt, data=.))

models2 %>% 
  map(summary) %>% 
  map_dbl(~.$r.squared)
#>         4         6         8 
#> 0.5086326 0.4645102 0.4229655

Created on 2019-02-26 by the reprex package (v0.2.1)

Upvotes: 1

Related Questions