1.1 Where Does a Query Go?

Source: Dev.to

PostgreSQL Internals – Chapter 1: Query Processing

When a client sends a statement such as

SELECT * FROM users WHERE id = 1;

the SQL travels through a five‑stage pipeline inside the PostgreSQL backend before a result row is returned.

1. Backend entry and dispatch

The backend receives the client message and decides which processing path to follow.

2. Parser and Analyzer

The SQL text is parsed, and the system catalog is consulted to give the query meaning.

3. Rewriter

The RULE system expands views, injects row‑level‑security (RLS) policies, and otherwise transforms the query tree.

4. Planner

Using a cost model, PostgreSQL explores possible execution paths and selects the most efficient one.

5. Executor

The chosen plan is walked, tuples are pulled up, and the results are sent back to the client.

---

Chapter 1.1 – Entry Point

Chapter 1.1 focuses on the first stage of the pipeline—the backend’s entry point. It is divided into three sections:

1.1.1 Life of a Query

A single diagram that compresses all five stages, providing a map for the rest of the book.

1.1.2 Simple vs Extended

Explores the semantic differences between the Simple Query protocol and the Extended Query protocol.

1.1.3 Optimizable vs Utility

Shows how optimizable statements (e.g., SELECT, INSERT, UPDATE, DELETE) follow a different path from utility commands (e.g., CREATE, VACUUM).

After completing this chapter, you should understand how the backend’s main loop receives a client message and dispatches it to the appropriate function.