Alert Using Same Key-Card Three or More Times in a One Hour Period

Alert Using Same Key-Card Three or More Times in a One Hour Period

What is this problem about?

The "Alert Using Same Key-Card Three or More Times in a One Hour Period interview question" is a data-processing challenge that mimics real-world security monitoring. In this scenario, you are given a list of names and the times at which they used their key-cards to enter a building. The goal is to identify users who triggered an alert by using their card three or more times within any rolling 60-minute window. This problem tests your ability to handle time-series data, sort events, and efficiently check constraints across a sliding timeframe.

Why is this asked in interviews?

Companies like Snowflake and Karat ask this "Alert Using Same Key-Card Three or More Times in a One Hour Period coding problem" because it evaluates a candidate's proficiency with hash tables and sorting. It also reflects common tasks in system design, such as rate limiting, fraud detection, and logging analysis. Understanding how to group events by a unique identifier (like a username) and then analyze those events chronologically is a fundamental skill for any backend or data engineer.

Algorithmic pattern used

This problem primarily uses the Hash Table and Sorting patterns. First, you need to group all access times by the specific user using a hash map. Once the data is grouped, the "Array interview pattern" comes into play: you sort the times for each user. Finally, a Sliding Window or a simple linear scan over the sorted times allows you to check if the difference between the first and third access in any three-consecutive-access sequence is 60 minutes or less.

Example explanation

Imagine you have a user "Alice" with the following access times: ["10:00", "10:20", "10:55", "11:10"].

1. Convert these to minutes: [600, 620, 655, 670].
2. Group and sort: For Alice, the sorted times are [600, 620, 655, 670].
3. Check Windows:
   • Window 1: 600 to 655. The difference is 55 minutes. Since there are 3 accesses (10:00, 10:20, 10:55) within 60 minutes, Alice triggers an alert.
   • Even if we checked the next window (620 to 670), she is already flagged.

Common mistakes candidates make

• Not converting time correctly: Failing to convert "HH:MM" into a total number of minutes makes comparisons difficult and error-prone.
• Ignoring the sorting step: You cannot check the "one-hour period" efficiently if the access times are processed out of order.
• Off-by-one errors: The window is "within one hour," meaning a difference of exactly 60 minutes (e.g., 10:00 to 11:00) is inclusive.
• Duplicate alerts: Forgetting to return only unique names in alphabetical order, as requested by most versions of this problem.

Interview preparation tip

When dealing with time-based problems, always normalize your data as early as possible. Converting "HH:MM" to a single integer representing minutes from the start of the day simplifies all subsequent arithmetic. Practice grouping data using maps and consider the edge cases where an alert is triggered exactly at the 60-minute mark.