Forum Discussion

StephenMadison's avatar
StephenMadison
Qrew Cadet
7 years ago

Median of Field Entries

I need to be able to calculate the median of numeric fields.  I know that we aren't able to do it natively so I'm hoping that someone from the community has had to do this before and can help.  There's an old post that references an older java site but its no longer available.

Basically, my organization only calculates the time between 2 dates and then gathers all those durations and finds the median.  This is the only metrics they go by. 
  • On average how many numbers are you talking about finding the medium of? Because this would be a VERY complicated process to do in Quickbase without using javascript.

    A type of vanilla bubble sorting. Are you trying not to use javascript? 




    • StephenMadison's avatar
      StephenMadison
      Qrew Cadet
      It will vary as it will need to be scalable the more entries we put into it.  And No i would LOVE to use javascript, I've been practicing with onload on other apps and I can get that function to work but with the median, I would need to know what to put and where.
    • _anomDiebolt_'s avatar
      _anomDiebolt_
      Qrew Elite
      Apparently the Insert-sort with Romanian folk dance is the fastest algorithm at four minutes and three seconds so I would go with that rather than the Bubble-sort with Hungarian ("Cs�ng�") folk dance which comes in at five minutes and fifteen seconds.
    • StephenMadison's avatar
      StephenMadison
      Qrew Cadet
      Hey Dan, i'm not sure if you're making some sort of joke but welcome to the party.  :)
    • _anomDiebolt_'s avatar
      _anomDiebolt_
      Qrew Elite
      I am a pretty serious guy and would never joke around especially when dealing with mathematical topics or algorithmic folk dancers.

  • I�m pretty swamped at work right now, but I�ll write the code out for you in awhile if you still need.
  • Here is a JavaScript function getMedian that takes (1) a table dbid(2) a query string, and (3) a fid of a numeric field and returns a promise that resolves to the median:

    function getMedian(dbid, query, fid) {
      return Promise.resolve(
        $.get(dbid, {
          act: "API_DoQuery",
          query,
          clist: fid,
          fmt: "structured"
        })
      ).then(function(xml) {
        var data = $('qdbapi table records record f[id=${fid}]', xml).map(function() {
          return parseFloat($(this).text());
        });
        data.sort((x, y) => x - y);
        var ilm = Math.floor((data.length - 1) / 2);
        var ihm = Math.ceil((data.length - 1) / 2);
        var median = (data[ilm] + data[ihm]) / 2;
        return median;
      })
    }

    You call it like this:
    var dbid = "bnxssrg9s";
    var dbidTable = "bnzd357bm";
    var apptoken = "cnurzq5cgchgjgdsdiup6cgkq6ct";
    $.ajaxSetup({data: {apptoken}});
    var query = "{3.GT.0}";
    var fid = "8";
    getMedian(dbidTable, query, fid)
      .then(function(median) {
        console.log(median);
      });

    When called on a sample application with this numeric data:

    0: 60000
    1: 65000
    2: 70000
    3: 75000
    4: 85000
    5: 155000
    6: 160000
    7: 200000

    it correctly calculates the median as 80000 (average of midpoints 75000 and 85000).

    The organization of this script can calculate any mean, metric or norm of a list of numeric data.

    I created this hypothetical GUI interface so QuickBase developers can hardcode the most popular means, metrics and norm functions: