Use Position Values to Keep Values in Matrix

Hi all, this is a little confusing to explain so I'll write down what I'm trying to do and then give a shorter example that may be helpful for generating the code.

Long version:

I have a matrix called Filling of dimensions 10834 x 500, which indicates how much sediment is deposited within a river channel. A separate matrix, Volume, has the same dimensions as Filling (10834 x 500) but indicate when sediment load is below or exceeds the volume of the channel. When the load is below the channel volume, the value of the cell within the matrix Volume is positive but when the load is higher than the channel volume, the value within the cell of matrix Volume is negative. Based on the positive and negative values within the matrix Volume, I then created an array called Position that indicates which column has the switch from positive to negative values for every row in the matrix Volume. This gives the array Position dimensions of 10834 x 1.

I now want to take the values with the array Position (10834 x 1) and apply it to the matrix Filling (10834 x 500) so that the value within Position results in keeping the value of the corresponding column while making the other values within the row 0. I would like to call the resulting matrix Filled and have it be of dimensions 10834 x 500. Each row within the matrix Filled should thus only have one value with all other columns of that row being 0. I've tried doing this through indexing but each possible combination I've tried doesn't result in what I'm wanting.

Short version:

To make this slightly easier, let's say the first matrix Filling is of size 5 x 3:

Filling=[1 2 3; 4 5 6; 7 8 9; 1 2 3; 4 5 6]

The other matrix, Volume, is also of size 5 x 3, but contains positive and negative values:

Volume=[1 -1 -2; 1 -1 -2; 2 1 -1; 3 2 -1; 1 -1 -2]

The array, Position, is 5 x 1 and indicates which column within Volume has the switch from positive to negative values:

Position=[2;2;3;3;2]

I now want to apply the array Position to the matrix Filling so that the values within the column that corresponds to the value within Position is kept while all other values become 0. I want to call this new matrix Filled and the resulting dimensions should be 5 x 3. The end result of Filled should look something like this:

Filled=[0 2 0; 0 5 0; 0 0 9; 0 0 3; 0 5 0]

To repeat, how can I end up with the new matrix called Filled by using the existing matrix called Filling and the existing array called Position? Thanks in advance!

Answer 1

[David Hill]

Filled=Filling.*(cumsum(Volume<0,2)==1);

[Derrick Vaughn]

Hi David, thanks for that code! It worked for what I described above.

There are a few rows within Volume where the values do not switch from positive to negative (values remain positive). Is there any way the code can be adjusted so that the last value in the last column of those rows get used (if the last value is not used, then every value with each of those rows would be 0). I imagine this would need to be done using an if/then statement?

[David Hill]

Just set the last column of Volume to -1. Then run the code above.

Volume(:,end)=-1;