This is a great question and test of Tableau Prep!
My solution is similar to yours in that I did the same aggregations to get the distinct sets of dates and Part / FMC combos. I also did the join on 1 = 1 to end up with the full set of possible values. But from there, I think all that's needed is to left join in the original data set - because you have the full set of data and just need to know whether you have a match in the original or if you should have a NULL (or possibly 0) value for the "empty" rows.
Here's what I ended up with:
Here's the result of my final step:
I've attached the packaged flow (.tflx) so you can take a look! It's always fun to see how other people approach the same problem.
(Also, shameless plug, but I'm going to be presenting a similar technique at TC18: https://tc18.tableau.com/learn/sessions?search=dragon&apply_filter=yes )
Data Oragami.tflx 9.0 KB
I like the simplification of the flow. Good luck taming the dragon at TC18, might see you there.