Thanks to Henry and all the others who posted possible solutions! Henry wrote: > I don't get what your 'offset' field is
My use of this verb might have more general applications down the road, but my immediate need was for a sort of "month odometer" that could start with any month of a year (not just January). The "offset" was the offset from the beginning, so that Jan=1, Feb=2, and so on, making 9 be the value for starting with September. I didn't mention months because I wanted a general solution to the problem, not a specific one that would work only with months. and Ric wrote: > I'm not sure what the point of giving a datalist is though since the > only thing that gets used from it is the number of items in the > list. I used the i.<n> approach in the problem since that's pretty much the standard way of signifying data in sample programs. In hindsight, I probably should NOT have used that approach, since that approach could actually have made the solution less general since it might work only for data in consistent serial order. Although datalist=. 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18... is ONE set of data which should work with the proposed algorithm, the proposed solution(s) should ALSO work with something like this: datalist=. 0 3 4 10 14 15 16 17 21 26 27 28... I know that Henry's solution gives the correct values for this datalist, and the others might as well (I didn't test them). My next step is to handle the turnover of years for each 12 months as a formua/algorithm that operates on the whole dataset at once (at the same time as this current algorithm). I prefer trying things myself first, but you may see another question to the forum if I'm stumped. I have yet another question, but since it's a different topic, I'll pose it separately. Thanks again for the solutions! Harvey ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
