Sliding Window Pattern

When to use Sliding Window

• The problem asks for a contiguous subarray or substring that satisfies a property (max sum, longest, shortest, count).
• Brute force would re-scan the same elements for every starting index.
• The property of the window can be maintained incrementally as one element enters and one leaves (sum, character count, distinct count).
• The window size is fixed (k) or variable based on a constraint that you can detect at the edges.
• You can answer the question by looking at the current window only, without backtracking.

Template

function variableSlidingWindow(s, constraint) {
  let left = 0;
  let best = 0;
  const state = new Map();
  for (let right = 0; right < s.length; right++) {
    state.set(s[right], (state.get(s[right]) || 0) + 1);
    while (!isValid(state, constraint)) {
      state.set(s[left], state.get(s[left]) - 1);
      if (state.get(s[left]) === 0) state.delete(s[left]);
      left++;
    }
    best = Math.max(best, right - left + 1);
  }
  return best;
}

Common mistakes

• Shrinking the window from the wrong side, breaking the contiguous-range invariant.
• Updating the running aggregate (sum, count) before deciding whether to record a new best, leading to off-by-one window sizes.
• Using a `for` loop on left instead of a `while`, which fails when multiple elements must leave for the window to become valid again.
• Forgetting to delete a key from the hash map when its count hits zero, which inflates the distinct-character count.
• Confusing fixed-size with variable-size sliding window templates and mixing the two control flows.