在 R 的特定列中用随机小数替换 NA

Question

I am trying to replace NA s with random decimals in a particular column in R. However, R generates random decimals with the same trailing fraction and just changes the part before the decimal. The following are the methods I tried:

df_LT$ATC[is.na(df_LT$ATC)]  <- sample(seq(10.2354897,23.78954214), size=sum(is.na(df_LT$ATC)), replace=T)

dplyr

df_LT <-  df_LT %>%mutate_at(vars(df_LT$ATC), ~replace_na(., sample(10.2354897:23.78954214, size=sum(is.na(ATC)), replace=T)))

Data looks as below

    A        ATC
    1        11.2356879
    2        42.58974164
    3            NA
    4        34.25382343
    5             NA 

Now, wherever there is a NA in the ATC column I want to add a decimal like the others but in the range 10:23. Hope this explanation will help. I may be missing something very obvious. Thanks for the help in advance.

Answer 1

You are using seq or the colon operator : to create your samples, which means you are sampling from following sequence:

seq(10.2354897, 23.78954214)
# [1] 10.23549 11.23549 12.23549 13.23549 14.23549 ....

So the starting value is increased by 1 in each step, leaving the numbers after the decimal points fixed.

If you want to sample random number within the range of these two limits you can do:

runif(n = 1, min = 10.2354897, max = 23.78954214)

So for your example this translates to:

df_LT$ATC[is.na(df_LT$ATC)] <- 
  runif(n = sum(is.na(df_LT$ATC)), 10.2354897, 23.78954214)

If you want to add a condition you can do:

df_LT$ATC <- 
  ifelse(is.na(df_LT$ATC) & df_LT$A == 3, 
         runif(n = nrow(df_LT), 10.2354897, 23.78954214), 
         df_LT$ATC)

This checks whether ATC is missing and also whether A is equal to 3. If this is fulfille the missing value is replaced with a random number, otherwise the original value (missin or not) is returned.