>Hi Ed,
>
>>>Am I dreaming? Any ideas welcome.
>>
>>With no rule for resolving processing order or limits on coverage, yep; sleepy time...
>
>I was hoping for some fancy iif() or self-join that wasn't obvious to me. Or even a better way to handle the second pass than a scan...endscan. Hmmm.

The problem is the order of process and accumulation of results must be done line by line clustered by incident and depleting liability until either the coverage or the cost of the incident is exhausted; any further coverages are ignored. I don't see a strong alternative to sequential processing, especially since there is no single fixed rule for application of coverage.