1 Reply Latest reply on Nov 15, 2016 11:18 AM by Shinichiro Murakami

    Filtering duplicate records by time difference

    Anthony Paquet

      Hello everyone,

       

      My phone system seems to generate duplicates somewhat randomly in its records. This is a problem for me as different visualizations will display an exaggerated result (e.g. total number of calls twice what it's supposed to be).

       

      I went with a first attempt using COUNTD(Time of call). This filtered out some duplicates, but the fact is that some duplicates are recorded with one second in difference and thus "escape" the COUNTD (see screenshot below). It would also potentially eliminate some legitimate records.

       

       

      I have thought about counting the records while considering all calls from the same number and recorded within the same X seconds as one (e.g. 5 seconds) but I have absolutely no idea of how to do this (or if this is even possible). I have previously used IFs but I am not sure if they apply to this context.

       

      Thanks for your help!