If the value in the second column of the matrix is greater than the specified constant, then you need to replace the row containing this value by the number of rows equal to quotient of division of this number by constant + 1. In the first rows, the value in the second column must be equal to a constant, and in the last row the value in the second column must be equal to a remainder of the integer division of the number by a constant.

On the left is the original matrix, on the right is what should be obtained.

That's a rough implementation, although I probably misunderstood some part.

It should be something like this but probably there is a better decision

Seems I missed the physical meaning, this should be a better implementation:



split layers.sm (10 KiB) downloaded 32 time(s).

It is magic! Thank you!
Can you do the same whith distance units?

Wrote

It is magic! Thank you!
Can you do the same whith distance units?

There's not much to change, just add the units to:

• threshold
  
• const
  
• rem, here you have to reintroduce the unit (automated by UnitsOf() or directly)

Hint: You can get a submatrix like this:

subMatrix = fullMatrix(1:someRow, : );
subMatrix = fullMatrix(row1:row2, : );
subMatrix = fullMatrix(someRow:end, : ); % End has a special meaning

(pick one), and you can concatenate matrices vertically like this:
tallMatrix = [shortMatrix1; shortMatrix2]; % Use as many as you want
They all have to have the same number of columns of course.

I hope this helps, thanks!