r - extract regex from a column and replace

Question

I am hoping to read a file and process it in R. It follows this format:

tibble(row=8,name="Shawn",state="Arizona", age=45,
comments="father=Alex;NSM;GNO;One:0.9995,0.13|Two:0.9991,0.55|Three:0.9996,0.33|Four:0.9986,0.22|Five:0.9987,0.22")

# A tibble: 1 x 5
        row name  state     age comment                                                                          
      <dbl> <chr> <chr>   <dbl> <chr>                                                                             
    1     8 Shawn Arizona    45 father=Alex;NSM;GNO;One:0.9995,0.13|Two:0.9991,0.55|Three:0.9996,0.33|Four:0.9986…

I would like to replace the last column with

comment
One:0.9995|Two:0.9991|Three:0.9996|Four:0.9986|Five:0.9987

and then subtract the floating number from 1:

comment
One:0.0005|Two:0.0009|Three:0.0004|Four:0.0014|Five:0.0013

Answer 1

This function is not particularly performant or elegant, but it will give you exactly the output you asked for. It is not vectorized, but you could use sapply to fix that.

sample <- "father=Alex;NSM;GNO;One:0.9995,0.13|Two:0.9991,0.55|Three:0.9996,0.33|Four:0.9986,0.22|Five:0.9987,0.22"

extract_numbers <- function(x) {
  x <- sub(".*;", "", x)
  x <- gsub(",([0-9.]*)|", "", x)
  x <- strsplit(x, "|", fixed = TRUE)[[1]]
  do.call(
    "paste", 
    c(
      lapply(
        strsplit(x, ":"), 
        function(y) paste(y[1], sprintf("%6.4f", 1 - as.numeric(y[2])), sep = ":")
      ),
      sep = "|"
    )
  )
}

extract_numbers(sample)
# [1] "One:0.0005|Two:0.0009|Three:0.0004|Four:0.0014|Five:0.0013"