I tried to use LAG functions to score each row 1 for a new start row and 0 for others. So column has values of 1 and 0 eg 1,0,0,1,1,0,1 then uses SUM aggregate function to add all values together so becomes 1,1,1,2,3,3,4. Then use these values to group row

v2

v1

Using LAG and LEAD functions to get start admission date and upper bound admission date respectively, then joining on original table to get amounts.

Logic wasn't so bad but the formatting was a real struggle.

It's all a bit messy really. I think I can improve on it.

It's all a bit messy really. I think I can improve on it.

Well, three things ocurred to me, firstly that two line item records with the same price must share the same TaxCode. Also, there is no need to recursiively determine combinations that make up the highest Amount because it must be all unallocated line item

Well, three things ocurred to me, firstly that two line item records with the same price must share the same TaxCode. Also, there is no need to recursiively determine combinations that make up the highest Amount because it must be all unallocated line item