Phase Three, Physical Planning: Choosing How to Execute

Logical optimization produces an optimized logical plan that says what to compute but not how. Physical planning is where Catalyst decides the how: it generates one or more physical plans, concrete strategies for actually running each operation, and chooses among them. This is the phase where a groupBy becomes a specific aggregation strategy and, most importantly, where a join becomes a specific join algorithm. Focus on the join strategy choice, where physical planning has its largest effect on performance. The same logical join can be executed as a broadcast hash join, where a small side is shipped to every executor and no big-side shuffle happens, or as a sort-merge join, where both sides are shuffled and sorted by key, or as a shuffle hash join. Catalyst picks based on its estimate of t

About This Interactive Section

This section is part of the Inside Catalyst: The Four Phases lesson on DataDriven, a free data engineering interview prep platform. Each section includes explanations, worked examples, and hands-on code challenges that execute in real time. SQL queries run against a live PostgreSQL database. Python runs in a sandboxed Docker container. Data modeling problems validate against interactive schema canvases. All content is framed around what data engineering interviewers actually test at companies like Meta, Google, Amazon, Netflix, Stripe, and Databricks.

How DataDriven Lessons Work

DataDriven combines four interview rounds (SQL, Python, Data Modeling, Pipeline Architecture) with adaptive difficulty and spaced repetition. Easy problems get harder as you improve. Weak concepts resurface until you master them. Your readiness score tracks progress across every topic interviewers test. Every lesson section ends with problems you solve by writing and running real code, not by picking multiple-choice answers.