6 Replies Latest reply on Jun 24, 2020 5:44 PM by Darren Wrigley

    Using Data & MetaData Rules Conjunctively?

    Thomas Ferguson Active Member

      Hello, fellow EDC users!

       

      We're starting to push on data discovery in earnest in our use of EDC and I hope that someone may have some insight into the following:

       

      Our initial data discovery is seeded by using both metadata and data rules in EDC Catalog Administrator. Using them together leads to a disjunctive behavior, i.e., if I have a metadata rule FOO and a data rule BAR attached to a data domain BAZ, the default is to classify anything that has FOO in the column name or BAR among the data with the data domain BAZ.

       

      That's extremely permissive; in many cases (actually, probably all cases), I'd like to use the two in tandem---i.e., conjunctively---in order to get assignments to more fine-grained domains. For example, if I have a data domain "CEO Phone Number", a plausible pair of metadata and data rules would be:

       

      • Check if the string "CEO" is in the column name,and
      • Check that the data is of the form NNN-NNNN

       

      (Yes, there are phone numbers not of that form; this is just a toy example, I assure you!) The desired behavior is that fields that both are telephone numbers in form and have a syntactic indicator that they belong to a CEO will be classified as "CEO Phone Number."

       

      In practice, all phone numbers and anything with CEO in the column name (e.g. "CEO_MIDDLE_NAME") will be suggested as "CEO Phone Numbers."

       

      Is there a mode or something that we can switch on to get the desired behavior from the tool? Any thoughts on how to approach this type of use-case?

       

      Thanks in advance!