Greg Low covers a common scenario:
But another area that I see very few handling well, is the data that is missing, rather than just the data that is present. There’s a huge difference between data that arrived, and is odd, and data that just didn’t arrive at all.
One tool that’s great at working with streams of data is the Real Time Intelligence workload for Microsoft Fabric. And it’s also great at working with data that is missing from those streams.
Greg covers some of the scenarios around missing data, though not a lot on the process to fix them.