Output loop Kolmogorov Smirnov Test (ks.test) in dataframe

Question

I would like to incorporate the output of a ks.test recorded in a loop into a data frame or file, instead of printing the outputs of 1155 tests in the console... :-).

column_equality_stats = function(data, lab_stats1, lab_stats2, min_count=100) {
  for(i in 1:length(lab_stats1)) {
    lab_testcodes_1 = lab_stats1[i]
    lab_testcodes_2 = lab_stats2[i]
    equal_columns <- filter(data, lab_testcode==lab_testcodes_1 | lab_testcode==lab_testcodes_2)
    col1 <- equal_columns[equal_columns$lab_testcode==lab_testcodes_1, 'lab_result']
    col2 <- equal_columns[equal_columns$lab_testcode==lab_testcodes_2, 'lab_result']
    if(sum(!is.na(col1))>min_count && sum(!is.na(col2))>min_count){
      stats <- ks.test(col1, col2)
      print(stats)

    }
  }
}

I would like to have a data.frame with the following columns: the names of col1 and col2 (the equation values), the p-value and the D-value.

Utopian data frame

Thank you very much in advance!!

Answer 1

This is as far as I can go. Please edit your question and add a reproducible example.

I need to know how your data is defined. I can't run your code! :-)

Anyway, just create a dataframe for every run of the loop and bind them together.

map_dfr of the purrr package does it for you.

library(purrr)

.column_equality_stats <- function(i, data, lab_stats1, lab_stats2, min_count = 100){

  lab_testcodes_1 <- lab_stats1[i]
  lab_testcodes_2 <- lab_stats2[i]
  equal_columns <- filter(data, lab_testcode==lab_testcodes_1 | lab_testcode==lab_testcodes_2)
  col1 <- equal_columns[equal_columns$lab_testcode==lab_testcodes_1, 'lab_result']
  col2 <- equal_columns[equal_columns$lab_testcode==lab_testcodes_2, 'lab_result']

  if(sum(!is.na(col1))>min_count && sum(!is.na(col2))>min_count){

    stats <- ks.test(col1, col2)
    res <- data.frame(col1 = lab_testcodes_1,
                      col2 = lab_testcodes_2,,
                      pvalue = stats$p.value,
                      dvalue = stats$statistics)

  } else {res <- data.frame()}

  res

}

column_equality_stats <- function(data, lab_stats1, lab_stats2, min_count=100) {

  map_dfr(seq_along(lab_stats1), 
          .column_equality_stats,
          data       = data,
          lab_stats1 = lab_stats1,
          lab_stats2 = lab_stats2,
          min_count  = min_count)

}