Feature Quality by Source

A medium SQL interview practice problem on DataDriven. Write and execute real sql code with instant grading.

Domain

SQL

Difficulty

medium

Seniority

mid, senior

Problem

Find sources in the ML feature store that have both high-null features (null percentage above 20) and low-null features (null percentage below 2). For each qualifying source, count how many different features fall into each bucket. Include sources even if one bucket is empty.

Practice This Problem

Solve this SQL problem with real code execution. DataDriven runs your SQL queries against a live PostgreSQL database and grades it instantly.