What is a decorator factory (a decorator with arguments)?

Q: What is a decorator factory (a decorator with arguments)?

**Target Audience:** Developers who can write a plain decorator and want configurable ones **Difficulty:** Medium **Python version:** behaviour checked on CPython 3.11 (python3.11 and the in-page Pyodide 0.25.1 runtime) > **How to read this doc:** Every idea is first explained in plain words. When a term has a precise name an interviewer expects to hear, it is introduced as a 📌 **Interview term** callout. The Quick Glossary at the end lists all of them. ## 1. Why This Even Matters — A Story First A stamp shop sells one stamp that says "PAID". Useful, but you want "PAID 3 TIMES", "PAID IN EUR", "PAID BY CARD". The shop does not carve a new stamp by hand for every order; it has a **stamp-making machine**. You type the text, the machine produces a stamp, and then you use that stamp on your papers. A plain decorator is the fixed stamp: `@timer`. A **decorator factory** is the machine: you give it settings, `repeat(3)` or `retry(times=5, on=(ConnectionError,))`, and it hands back a rea

The problem

Question presented to candidate: "Writing @timer is easy. Now I want @retry(times=3) or @repeat(3), where the decorator takes its own arguments. How do you write that, and what does Python actually do with the @repeat(3) line?"

What a strong answer should cover:

@repeat(3) first calls repeat(3), which must return a decorator; that decorator is then applied: f = repeat(3)(f).
Three layers: the factory takes the options, the decorator takes the function, the wrapper takes the call arguments; options live in closure cells.
Timing: the factory and the decorator both run once at definition time; only the wrapper runs per call.
functools.wraps(func) goes on the innermost wrapper.
The classic bug of using a factory without parentheses (@repeat), and how to support both @deco and @deco(opt=...) with func=None plus keyword-only options.
Alternatives: a class whose __init__ takes the options and __call__ takes the function; standard-library examples such as functools.lru_cache(maxsize=...).

Code / implementation expected: Yes — a parameterised decorator such as repeat(n) or retry(times, exceptions), written live.

decorators
factory

Try it yourself

Sketch your approach before you read the solution. That is what the interview tests.

Stuck? Get a hint.

A short nudge toward the approach. It will not give the answer away.

Solution

Target Audience: Developers who can write a plain decorator and want configurable ones Difficulty: Medium Python version: behaviour checked on CPython 3.11 (python3.11 and the in-page Pyodide 0.25.1 runtime)

How to read this doc: Every idea is first explained in plain words. When a term has a precise name an interviewer expects to hear, it is introduced as a 📌 Interview term callout. The Quick Glossary at the end lists all of them.

1. Why This Even Matters — A Story First

A stamp shop sells one stamp that says "PAID". Useful, but you want "PAID 3 TIMES", "PAID IN EUR", "PAID BY CARD". The shop does not carve a new stamp by hand for every order; it has a stamp-making machine. You type the text, the machine produces a stamp, and then you use that stamp on your papers.

A plain decorator is the fixed stamp: @timer. A decorator factory is the machine: y

The solution is hidden so you can try first

10 min read · 12 sections

Run the code

Edit the code and run it here. Each console line with an expected value becomes a test.

PythonThree layers, when each runs, reuse, and a retry factory

Practice

PythonThe missing-parentheses bug, an optional-argument decorator and a class factory