External Merge Sort — Sorting Data Bigger Than RAM

Here's a question that separates junior from mid-level DE candidates: 'You have a 500GB CSV file. Sort it by user_id. You have 8GB of RAM.' The right answer is external merge sort. This two-phase algorithm is the foundation of how databases, MapReduce, and Spark handle large-scale sorting — understanding it shows you connect algorithms to real infrastructure. Full Implementation Chunk Size Tuning