The new "Raster rank" algorithm takes a collection of rasters as input and creates an output with cell values matching the rank order of a list of cell values taken from the input collection of rasters.
For example, if the list of input values for a given cell is [10,20,30,40] and the rank is 2, the output value will be 20.
The algorithm also allows for a negative index rank. Given the same list of input values, [10,20,30,40], and using the rank value -2, the output value would be 30.
Finally, the algorithm offers to nodata handling behavior. By default, nodata values are excluded from the list, and a given output cell will only be set to nodata if the rank is greater than the values list size. Users can also pick an alternative behavior, whereas one or more nodata values for a given cell in the input layers will result in a nodata value for that output cell.