My data is similar to the format below. Each user should have an event_1 and an Event_4 record for every day from 2012-01-03 to 2012-01-16. I am trying to classify users into 4 groups: "Complete", "No Event_1", "No_Event_2", "No Event_1 and No Event_2". I cannot see how to rind the missing records when they don't have either for the date.
Any direction would be helpful.
There are few ways to approach this,
One route is to generate a list of all Date-UserID combinations of interest, using this list as the primary data source, and then data blend in this sample data to produce a final result.