we have a number of datasets that involve date ranges. you have participantId, startDate and endDate. a null end date means that this date range has not ended.

there's a number of instances where we group records to do things like 'find the last date when the animal had a roommate', 'last co-assignment', etc. both min(endDate) and max(endDate) will never report the null value. these functions return either the highest or lowest non-null value. null is completely ignored.

at least for this particular application, it would be nice if null were treated as the max() of a set of dates. it already sorts that way. unfortunately it seems that this is just how postgres handles dates. i ended up doing max(coalesce(endDate, now()) in other to handle the null values. it works, but is not great. is there a better way? thanks for the help.