Custom date of birth - Matching


As part of our matching process we break date of birth into day / month and year and then have a custom rule which matches on 2 out of the 3 variables being correct. When combined with other data aspects such as surname and address we find this gives us better results.

DateCustom.Close={(DateDay.Exact & DateMonth.Exact) | (DateMonth.Exact & DateYear.Exact) | (DateDay.Exact & DateYear.Exact)}

I would however like to eliminate the possibility of father and son have being born on the same day | month and living in the same address by adding in a condition where the system checks to ascertain if the year of birth that has been given is + / - x number of years, possibly 5 years for arguments sake. This new variable would become DateYear.Close and would be included as

DateCustom.Close={(DateDay.Exact & DateMonth.Exact & DateYear.Close) | (DateMonth.Exact & DateYear.Exact) | (DateDay.Exact & DateYear.Exact)}

As such ...

John Smith 01/01/1980 would still match with

John Smith 01/01/1981 but would not match with

John Smith 01/01/1986

This is where I am trying to get to but I don't know if the system can have the + / - x number of years variable and if it can I certainly don't know how to code that in. Its not a Levenstein type 77% match and so I am a bit lost.

Anyone able to help or tell me it cant be done / point me to a better way of doing this. ?

Thanks again.


Best Answer


  • Chris DownerChris Downer Administrator

    Hi Steven

    Unfortunately we do not currently provide any data or number distance comparators in Find Duplicates. You can find lists of all our blocking algos and comparators here:

    For date we currently have:

    We do plan to add the following to Find Duplicates in the near future:

    • Data Differences +/- years, months, days
    • Number differences +/- a number range
    • Compare phonetics

    The only suggestion i have is to build a second match job that matches First name, last name, month of DoB and Day of DOB AND Cluster ID and then manually check the results. I can't imagine there will be many parent/child duplicates where the names and DOB are exactly the same!

  • Clinton JonesClinton Jones Experian Elite
    edited October 2019

    Just to add to this, of course there is a native function that supports a bunch of transformation options that would take a customer V1 sample data and extract the day month and year from the first order date and compare the first order date with today's date

    By using the rounded years option on compare date you could see if father and son with the same names and address have an age difference

  • Tanj JagpalTanj Jagpal Administrator

    @stevenmckinnon did this answer your question?

  • Thanks for this Danny ... this is going to take me a bit of time to read through and walk through the steps. Will try both options and see what brings back the highest results.

Sign In or Register to comment.