@Mike ,
Truly my pleasure - you’ve always got interesting problems, and do a great job framing the questions and providing the necessary data to make it easy to provide good support. Glad that got you what you needed.
The big take-away from this one is that 99.8% of the time when you might be tempted to use a bidirectional relationship, there is another way to do it using unidirectional relationships (sometimes that’s developing composite keys, de-normalizing dimension tables, using TREATAS, building a bridge table, etc.).
The only situation I’ve encountered where there simply wasn’t a good workaround is the Period Slicer configuration that eDNA Expert @JarrettM uses frequently, and that I’ve adopted in some of my reports as well.
However, this bidirectional relationship is isolated in the far corner of the data model, snowflaked off of the extended data table, where it can cause no ambiguity or other problems.
- Brian