Finding neighbours in a two-dimensional array
(pseudo-code)
row_limit = count(array);
if(row_limit > 0){
column_limit = count(array[0]);
for(x = max(0, i-1); x <= min(i+1, row_limit); x++){
for(y = max(0, j-1); y <= min(j+1, column_limit); y++){
if(x != i || y != j){
print array[x][y];
}
}
}
}
Of course, that takes almost as many lines as the original hard-coded solution, but with this one you can extend the "neighborhood" as much as you can (2-3 or more cells away)
I think Ben is correct in his approach, though I might reorder it, to possibly improve locality.
array[i-1][j-1]
array[i-1][j]
array[i-1][j+1]
array[i][j-1]
array[i][j+1]
array[i+1][j-1]
array[i+1][j]
array[i+1][j+1]
One trick to avoid bounds checking issues, is to make the array dimensions 2 larger than needed. So, a little matrix like this
3 1 4
1 5 9
2 6 5
is actually implemented as
0 0 0 0 0
0 3 1 4 0
0 1 5 9 0
0 2 6 5 0
0 0 0 0 0
then while summing, I can subscript from 1 to 3 in both dimensions, and the array references above are guaranteed to be valid, and have no effect on the final sum. I am assuming c, and zero based subscripts for the example
an alternative to @SebaGR, if your language supports this:
var deltas = { {x=-1, y=-1}, {x=0, y=-1}, {x=1, y=-1},
{x=-1, y=0}, {x=1, y=0},
{x=-1, y=1}, {x=0, y=1}, {x=1, y=1} };
foreach (var delta in deltas)
{
if (x+delta.x < 0 || x + delta.x >= array.GetLength(0) ||
y+delta.y < 0 || y + delta.y >= array.GetLength(1))
continue;
Console.WriteLine("{0}", array[x + delta.x, y + delta.y]);
}
Slight advantage in readability, possible performance if you can statically allocate the deltas.
Here is a working Javascript example from @seb original pseudo code:
function findingNeighbors(myArray, i, j) {
var rowLimit = myArray.length-1;
var columnLimit = myArray[0].length-1;
for(var x = Math.max(0, i-1); x <= Math.min(i+1, rowLimit); x++) {
for(var y = Math.max(0, j-1); y <= Math.min(j+1, columnLimit); y++) {
if(x !== i || y !== j) {
console.log(myArray[x][y]);
}
}
}
}