CODEWITHSUNDEEP
r
user17153595
user17153595

Reputation:

Fill NA values in a data frame with random values from a vector (without replacement)

Context: I'm trying to fill NA values for a column in my data called "Cholesterol" with a vector of sampled values, however, I couldn't find anything that could help with that. I've tried using replace_na, but it it not replacing the NA values

MRE:

Cholesterol data example

69 181 308 166 211 257 182  NA  NA  NA  NA  NA  NA  NA
[301]  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA
[331]  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA
[361]  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA
[391]  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA  NA 260 209 218 228
[421] 213  NA 236  NA  NA 267 166  NA  NA  NA  NA 220 177 236  NA  NA  NA  NA  NA  NA  NA  NA  NA 186 100 228  NA 171 230  NA
[451]  NA  NA 281  NA 203  NA  NA  NA  NA  NA 277  NA 233  NA  NA 240  NA  NA 153 224  NA  NA  NA 316  NA  NA 218  NA 311  NA
[481]  NA  NA 270  NA  NA 217 214 214 252 220 214 203  NA 339 216 276 458 241 384 297 248 308 208 227

Vector that will fill the NA values:

missing_values = sum(is.na(df$Cholesterol))
missing_values
    
# Set seed
set.seed(42)

fill_NA_values_cholesterol = sample(rnorm(n = missing_values, mean = mean(cholesterol_sem_valores_nulos, trim = 0.2), sd = mad(cholesterol_sem_valores_nulos)), size = missing_values)

The variable cholesterol_sem_valores_nulos is simply a different vector that only has filled valued (NAs are not present in this vector)

How could I make the code fill the NA values using the vector fill_NA_values_cholesterol? The amount of NA values present in df$Cholesterol is 172 (same length of the fill_NA_values_cholesterol)

Thank you in advance

Upvotes: 0

Views: 72

Answers (1)

jpiversen
jpiversen

Reputation: 3212

Here is an example, where I use purrr together with the rnorm() function you specified to replace the NA-values.


library(dplyr)

# Some example data
df <- tibble(
  Cholesterol = c(NA, 1:3, NA)
)

# I make this as a function to save some space underneath, but it is not
# necessary
draw_random_based_on <- function(x) {
  
  rnorm(
    n = 1, 
    mean = mean(x, trim = 0.2, na.rm = TRUE),
    sd = mad(x, na.rm = TRUE)
  )
  
}

# Under I add a new column - Cholesterol2 - where non missing values are the
# same as Cholesterol, but missing values are replaced by the random function
# you specified
df %>%
  mutate(
    Cholesterol2 = purrr::map_dbl(
      Cholesterol, 
      ~ifelse(
        is.na(.x), 
        draw_random_based_on(df$Cholesterol), 
        .x
      )
    )
  )

Upvotes: 0

Related Questions