Why persistence

Hub › Golang › Intermediate › Why persistence

Goal

Understand what the beginner endpoint is missing and what this tier fixes. No code on this page.

Prerequisites

• Golang Beginner — Serving JSON (the tier exit artifact)

What the beginner endpoint can't do

The beginner endpoint returns a hard-coded slice:

var items = []Item{
    {ID: 1, Name: "Notebook", Price: 4.0},
    {ID: 2, Name: "Pen", Price: 2.5},
    {ID: 3, Name: "Sticker pack", Price: 5.0},
}

That slice lives only in process memory. Two things make it a dead end for a real service:

1. No writes. You can't add, change, or remove an item through the API. The list is frozen at compile time.
2. No memory between restarts. Stop the server, start it again — the slate is blank. Any change you managed to make in code is gone the moment the binary exits.

Every production CRUD API replaces in-memory data with a database for exactly these two reasons.

What this tier adds

You will add a SQLite file (items.db) and a store layer that wraps it. By the end of page 08 you will have:

• GET /items — reads all rows from the database
• GET /items/{id} — reads one row by primary key
• POST /items — inserts a new row and returns it

The data survives restarts because SQLite writes it to disk. The test suite uses an in-memory SQLite connection so tests stay fast and leave no files behind.

Everything is pure Go — no cgo, no ORM, no framework. The only third-party dependency is a CGo-free SQLite driver.

Next → 02 Project setup