CODEWITHSUNDEEP
rfunctiondataframe
user12470302
user12470302

Reputation:

R Error: Multi-argument returns are not permitted, but there is no multi-argument

So I was tasked with helping my PI at my research lab come up with a simple function to help choose two random values from a dataframe whilst ignoring NA's. The data we work worth is based on self-report data so we have some areas left unanswered and we get NA's. We don't drop our NA's but during analysis for subscales we do need to select some values while ignoring NA's.

I have a simple test dataframe made and began working on my test function. I know it won't work as I need it to but I am running into a weird error. Here is my code.

rm(list = ls())                   ### clear all ###
dev.off(dev.list()["RStudioGD"])
options(scipen=999)

install.package("dplylr")
install.package("magrittr")
library("magrittr")
library("dplyr")

df<- data.frame("Var1"=c(1,7,8), "Var2"=c(NA, NA, 9), "Var3"=c(2, NA,10), "Var4"=c(3,5,NA), "Var5"=c(4,7,NA))

#select columns funciton, picks two random values from data frame not containing NA
select.col <- function(x,y) ##given df x select y columns
  df %>% 
    select_if(is.numeric) %>%
    return(z)

b <- select.col(df,2)

so, I'm trying to make a function which selects the second parameter as a random column from each row where NA is not selected.

I would want b to be a new data set and consist of these data in this format:


1, 2
2, 7
3, 9

(basically my data selected randomly without NA's...kind of like elementary x,y/input-output tables)

I ran into this snag and no matter how I rewrite this function I get the same error of "Multi-argument returns are not permitted."

Any ideas or advice?

EDIT: I fixed the formatting for how I want my data to look. Imagine it like a CSV; so in one column is one value and in another is the other value.

Upvotes: 0

Views: 2112

Answers (2)

Ronak Shah
Ronak Shah

Reputation: 389335

Do you need something like this?

select.col <- function(x,y) {
   nc <- y
   nr <- nrow(df)

   df %>%
     #select only numeric columns
     select_if(is.numeric) %>%
     #Convert data into a vector
     unlist %>%
     #Remove NA values
     na.omit %>%
     Select random nc * nr values
     sample(nc * nr) %>%
     #Convert it into matrix specifying number of rows and columns
     matrix(ncol = nc, nrow = nr) %>%
     #Convert into dataframe
     as.data.frame()
} 

select.col(df,2)

#  V1 V2
#1  5  1
#2  3 10
#3  8  2

Upvotes: 0

G5W
G5W

Reputation: 37661

When you write 

df %>% 
select_if(is.numeric) %>%
return(z)

It is the same as 

return(select_if(df, is.numeric), z)

return does not take two arguments, which is why you get the error message. Also, note that your return statement uses z, but z is not defined anywhere.

Upvotes: 1

Related Questions