Find minimum value greater than 0

Question

I have a data frame that contains numerical values 1:4 with some NA's. For each row, I would like to calculate the frequency (as a percentage) of the value with the fewest occurrences greater than 0.

Here is a sample data frame to work with.

    df = as.data.frame(rbind(c(1,2,1,2,2,2,2,1,NA,2),c(2,3,3,2,3,3,NA,2,NA,NA),c(4,1,NA,NA,NA,1,1,1,4,4),c(3,3,3,4,4,4,NA,4,3,4)))

      V1 V2 V3 V4 V5 V6 V7 V8 V9 V10
    1  1  2  1  2  2  2  2  1 NA   2
    2  2  3  3  2  3  3 NA  2 NA  NA
    3  4  1 NA NA NA  1  1  1  4   4
    4  3  3  3  4  4  4 NA  4  3   4

I have 2 points that I am struggling with. 1) finding the lowest frequency of a value greater than 0 and 2)applying the function to each row of my data frame. When I started working on this function I implemented it using the code below, but it did not appear to be applied to every row. My result for value.1, value.2, etc was the same for every row.

    Low_Freq = function(x){
      value.1 = sum(x==1, na.rm=TRUE) #count the number of 1's per row
      value.2 = sum(x==2, na.rm=TRUE) #count the number of 2's per row
      value.3 = sum(x==3, na.rm=TRUE) #count the number of 3's per row
      value.4 = sum(x==4, na.rm=TRUE) #count the number of 4's per row
      num.values = rowSums(!is.na(x), na.rm=TRUE) #count total number of non-NA values in each row

      #what is the minimum frequency value greater than 0 among value.1, value.2, value.3, and value.4 for EACH row?
      min.value.freq = min(cbind(value.1,value.2,value.3,value.4)) 

      out = min.value.freq/num.values #calculate the percentage of the minimum value for each row
    }

    df$Low_Freq = apply(df, 1, function(x))

Then I started using rowSums() to compute value.1, value.2, value.3, and value.4. This fixed my problem of counting value.1, value.2, etc for each row, however, I then had to apply the function without the use of apply() for it to run:

    Low_Freq = function(x){
      value.1 = rowSums(x==1, na.rm=TRUE) #count the number of 1's per row
      value.2 = rowSums(x==2, na.rm=TRUE) #count the number of 2's per row
      value.3 = rowSums(x==3, na.rm=TRUE) #count the number of 3's per row
      value.4 = rowSums(x==4, na.rm=TRUE) #count the number of 4's per row
      num.values = rowSums(!is.na(x), na.rm=TRUE) #count total number of non-NA values in each row

      #what is the minimum frequency value greater than 0 among value.1, value.2, value.3, and value.4 for EACH row?
      min.value.freq = min(cbind(value.1,value.2,value.3,value.4)) 

      out = min.value.freq/num.values #calculate the percentage of the minimum value for each row
    }

    df$Low_Freq = Low_Freq(df)

So the act of applying to each row then seemed to occur within the function itself. That's all fine and dandy, but when I go to make my final calculation which will be my output, I cannot figure out how to identify which of values 1, 2, 3, or 4 has the lowest frequency for each row. This value must be divided by the number of non-NA values for each row.

My desired result should look like this:

      V1 V2 V3 V4 V5 V6 V7 V8 V9 V10  Low_Freq
    1  1  2  1  2  2  2  2  1 NA   2 0.3333333
    2  2  3  3  2  3  3 NA  2 NA  NA 0.4285714
    3  4  1 NA NA NA  1  1  1  4   4 0.4285714
    4  3  3  3  4  4  4 NA  4  3   4 0.4444444

I feel like I am going in circles with this seemingly simple function. Any help would be appreciated.

Thank you.

Answer 1

The table function will return the frequency of each value that appears, ignoring NA values. Therefore, the min of the table result is the minimum frequency of a value that shows up in your row, and the sum is the number of non- NA values in your row.

Low_Freq = function(x){
  tab = table(x)
  return(min(tab) / sum(tab))
}
df$Low_Freq = apply(df, 1, Low_Freq)
df
#   V1 V2 V3 V4 V5 V6 V7 V8 V9 V10  Low_Freq
# 1  1  2  1  2  2  2  2  1 NA   2 0.3333333
# 2  2  3  3  2  3  3 NA  2 NA  NA 0.4285714
# 3  4  1 NA NA NA  1  1  1  4   4 0.4285714
# 4  3  3  3  4  4  4 NA  4  3   4 0.4444444

If you wanted to not use 5s for the numerator but to use them for the denominator, you could do:

df = as.data.frame(rbind(c(1,2,1,2,2,2,2,1,NA,2),c(2,3,3,2,3,3,NA,2,NA,NA),c(4,1,NA,NA,NA,1,1,1,4,4),c(3,3,3,4,4,4,5,4,3,4)))
Low_Freq = function(x){
  tab = table(x[x != 5])
  return(min(tab) / sum(!is.na(x)))
}
df$Low_Freq = apply(df, 1, Low_Freq)
df
#   V1 V2 V3 V4 V5 V6 V7 V8 V9 V10  Low_Freq
# 1  1  2  1  2  2  2  2  1 NA   2 0.3333333
# 2  2  3  3  2  3  3 NA  2 NA  NA 0.4285714
# 3  4  1 NA NA NA  1  1  1  4   4 0.4285714
# 4  3  3  3  4  4  4  5  4  3   4 0.4000000