CODEWITHSUNDEEP
rtidyversepurrr
user4687531
user4687531

Reputation: 1111

pmap not mapping across rows of dataframe

I'm trying to map a function to multiple inputs contained in rows of a dataframe using purrr::pmap. I've got a reprex below:

library(tidyverse)

# Create dummy data
e <- 0.3
f <- 1:3
g <- 1:3
df <- tidyr::crossing(e = e, f = f, g = g)

# Create a simple function to use
new_ratio <- function(x, y, z){
    return(x/(y + z))
}

# Apply pmap across rows of the dataframe `df` - not working as intended 
purrr::pmap(df, ~ new_ratio(x = e, y = f, z = g))
#> [[1]]
#> [1] 0.150 0.075 0.050
#> 
#> [[2]]
#> [1] 0.150 0.075 0.050
#> 
#> [[3]]
#> [1] 0.150 0.075 0.050
#> 
#> [[4]]
#> [1] 0.150 0.075 0.050
#> 
#> [[5]]
#> [1] 0.150 0.075 0.050
#> 
#> [[6]]
#> [1] 0.150 0.075 0.050
#> 
#> [[7]]
#> [1] 0.150 0.075 0.050
#> 
#> [[8]]
#> [1] 0.150 0.075 0.050
#> 
#> [[9]]
#> [1] 0.150 0.075 0.050

Created on 2018-07-16 by the reprex package (v0.2.0).

As it can be seen I really want to create a simple output row by row to give the ratio e/(f+g) in this simple example. The function keeps returning the same values for different inputs. In the first row, I would expect the output 0.3/(1 + 1) = 0.15, in the second row would expect 0.3/(1 + 2) = 0.1 etc

Could anyone please show how to correct for this? I am not quite understanding how to apply pmap from the documentation

Upvotes: 3

Views: 848

Answers (1)

akrun
akrun

Reputation: 886938

We can change the column names to the argument names and it should work

pmap_dbl(setNames(df, c('x', 'y', 'z')), new_ratio)
#[1] 0.150 0.100 0.075 0.100 0.075 0.060 0.075 0.060 0.050

Or if we intend to use ~, the arguments are .. followed by the number

pmap_dbl(df, ~ new_ratio(x = ..1, y= ..2, z = ..3))
#[1] 0.150 0.100 0.075 0.100 0.075 0.060 0.075 0.060 0.050

Upvotes: 1

Related Questions