Detect Rolling Window Anomalies | Dataford Interview Questions

Given a stream of transaction logs, write a function to detect anomalous patterns within a rolling time window.

Treat a transaction as anomalous when its inclusive rolling window contains at least two transactions, regardless of user. Return the zero-based indices of all anomalous transactions. Logs are ordered by nondecreasing timestamp.

The function accepts logs, a list of [timestamp, transaction_id] pairs, and window_seconds, a nonnegative integer. Return a list of indices.

Constraints

0 <= logs.length <= 10^4
Each log is a two-element list [timestamp, transaction_id]
0 <= timestamp <= 10^9
transaction_id is a non-empty string
0 <= window_seconds <= 10^9
timestamps are in nondecreasing order
A window is inclusive at both endpoints

Function Signature

def detect_anomalies(logs, window_seconds):

Your question is Detect Rolling Window Anomalies. Start with the requirements in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.

Problem

Given a stream of transaction logs, write a function to detect anomalous patterns within a rolling time window.

The function accepts logs, a list of [timestamp, transaction_id] pairs, and window_seconds, a nonnegative integer. Return a list of indices.

Constraints

0 <= logs.length <= 10^4
Each log is a two-element list [timestamp, transaction_id]
0 <= timestamp <= 10^9
transaction_id is a non-empty string
0 <= window_seconds <= 10^9
timestamps are in nondecreasing order
A window is inclusive at both endpoints

Function Signature

def detect_anomalies(logs, window_seconds):

Your question is Detect Rolling Window Anomalies. Start with the requirements in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.