Thanks for your help and the sample, Karl. Although the example is interesting, it seems designed for a slightly different problem than the one I'm considering. I'm looking for an QV expression that will accept my two columns of data (contact ID and event ID) and produce all the event pair combinations that a given contact ID has interacted with.
The data will allow my team to segment our customer base by similarity of attended events. I think QV could save me hours of off-line time if I could find an answer.
Here's an example with your data. You would have to handle 2 catalogs for events because one table can't have 2 columns EventID. If this isn't work you are looking for, please post an example of a report you would use for your analysis to understand it better.