Thank you for all your responses ! I tried, according to my knowledge (meaning if for me something doesnt work it is maybe my fault), all the solutions according to my data. A few key points : -proc transpose on a 800k table doesnt work ; the proc seems ok, but when i try to open the table, infinite lag -I specified (perhaps not enough) in my original message that the x of the Reachx can be another column (eg Reachx = Base*1.05, meaning there is Base, Reachx AND the colomn with the wanted number of row) ; in that way I cannot make the 2 proposed solutions from @mkeintz (I dont know how to tag) work with a variable column as Reachx (which change each row, see eg above) -The solution from @Ksharp work with a variable column, but it is unpractipable for a large table like mine ; for 30k rows it took 30sec, for 50k 1m30sec. To keep the continuity of the table it is probably possible tu cut the table into pieces and to make an overlap reconstruction happend after this solution for each cut, but it is not my priority now. But it is good to know ! I am very happy that at least one solution kinda work for me. I hope this will help others for future similar questions. I will accept the @Ksharp answer soon. Thank you 🙂
... View more