Longest Subarray with At Most Two Distinct Values

Given an integer array nums, return the maximum length of a contiguous subarray containing at most two distinct values.
The chosen subarray may contain either one or two distinct values.
Interview follow-up
The candidate proposed binary search on the answer with fixed-size window validation, and the interviewer asked them to implement that approach.
After reviewing the implementation, the interviewer pointed out that the problem can be solved directly with a sliding window.
The candidate explained the sliding-window approach and offered to implement it instead. The interviewer said another implementation was not necessary.

Function
longestSubarrayAtMostTwoDistinct(nums: int[]) → int

Examples
Example 1
nums = [1,2,1]
return = 3
The entire array contains only 1 and 2.
Example 2
nums = [0,1,2,2]
return = 3
The subarray [1,2,2] has length 3 and uses two distinct values.

The trick is a window with a frequency map. Move the right pointer, add nums[right] to the map, and while the map holds more than two keys, move the left pointer and decrement counts. Delete a key when its count hits zero. That delete step is the pitfall. If you only decrement and check the size of the map, a zero-count key still counts as distinct and you shrink too far. Track the best length after each right step. Edge cases: one element, all identical values, and exactly two values where the answer is the whole array. It's O(n) time and O(1) space since the map never exceeds three keys. The binary search on window size from the reported follow-up works but is more code and more bugs. Skip it. If you freeze mid-assessment, StealthCoder is the hedge that reads the problem and gives you the window pattern fast.

If you see this problem in your OA tomorrow, the play is to recognize the pattern in 30 seconds. StealthCoder buys you that recognition.