CODEWITHSUNDEEP
rlmpurrr
NelsonGon
NelsonGon

Reputation: 13319

Mapping additional arguments to purrr

I'm trying to map the function below to a nested dataframe using purrr. I've tried to look at the docs but just can't wrap my head around a way to pass additional arguments to the function.

Sample data:

library(dplyr)
library(purrr)
nested<-iris %>% 
  group_by(Species) %>% 
  nest()

Function to implement:

lm_mod<-function(df,xname,yname){
 xname<-deparse(substitute(xname))
 yname<-deparse(substitute(yname))
 f1<-as.formula(paste(yname,"~",xname))
 do.call(lm,list(data=quote(df),f1))
  }

One of several attempts:

map(nested$data,lm_mod(Sepal.Length,Petal.Length))

Error:

Error in stats::model.frame(formula = df, data = ~Petal.Length, drop.unused.levels = TRUE) : object 'Sepal.Length' not found

It appears that my formula is interpreted as the data. How can I best implement this?

Upvotes: 1

Views: 288

Answers (1)

NelsonGon
NelsonGon

Reputation: 13319

This fixed it: The problem was with my function definition that I later edited to explicitly state the data argument in the lm call and thanks to @MrFlick too for pointing that out:

lm_mod<-function(df,xname,yname){
 xname<-deparse(substitute(xname))
 yname<-deparse(substitute(yname))
 f1<-as.formula(paste(yname,"~",xname))
 do.call(lm,list(data=quote(df),f1))
  }

map(nested$data,lm_mod,Sepal.Length,Petal.Length)

Upvotes: 1

Related Questions