How would you handle null values and duplicate records in a large dataset of cloud resource usage telemetry for DigitalOcean?
