Java Practice Lab

Algorithms · 372 words · 2 minute read

Sliding Window in Java: Fixed Windows, Variable Windows, and Deques

By · Published 2026-09-04 · Updated 2026-09-04

A sliding window represents a contiguous range that moves through an array or String. It is useful when neighboring candidate ranges overlap heavily. Instead of rebuilding a sum or count for each range, remove the element leaving the window and add the element entering it. The hard part is not the syntax; it is defining exactly when the window is valid and when it must shrink.

Fixed-size windows

For a window of size k, build the first k values once. Every later move subtracts the value at right - k and adds the new right value. That turns a naive O(nk) calculation into O(n). Validate k before allocating output or dividing by it: k must be positive and no greater than the input length.

public class Main {
    static int bestSum(int[] values, int k) {
        int sum = 0;
        for (int i = 0; i < k; i++) sum += values[i];
        int best = sum;
        for (int right = k; right < values.length; right++) {
            sum += values[right] - values[right - k];
            best = Math.max(best, sum);
        }
        return best;
    }
    public static void main(String[] args) {
        System.out.println(bestSum(new int[] {2, 1, 5, 1, 3, 2}, 3));
    }
}

The output is 9, from the window 5, 1, 3. The loop never recomputes a window sum from scratch.

Variable windows and monotonic deques

For “longest range satisfying a rule,” grow right until the rule fails, then advance left until it holds again. The data structure holding the rule’s state might be a frequency map or count. For a maximum in every fixed window, a deque of candidate indices can do better than rescanning: discard expired indices from the front and smaller values from the back.

Common mistakes

  • Forgetting to remove the outgoing value before recording a new fixed window.
  • Allowing k equal to zero or greater than the input length.
  • Shrinking only once when a constraint may require repeated shrinking.
  • Storing deque values instead of indices, which makes expiration ambiguous.

Practice next

Start with maximum subarray sum to distinguish a dynamic subarray from a fixed window. Then solve sliding window maximum and compare its deque invariant with the simpler running-sum example. Continue to longest substring with k distinct characters when you are comfortable updating a frequency map as pointers move.

Continue with the Java practice path · Try code in the Java compiler