Level 3 exam — Node.js

Exam — fresh questions

12 questions to validate Express & databases: routing and middlewares, templates and static, Mongo/Mongoose, SQL and Prisma.
Answer everything, then click Check: your score and explanations will appear.

3.1 · Recall

1. In Express, you recognise middleware by…

Its server restart Its (req, res, next) signature in the pipeline Its colour in the terminal Its printing on paper

It acts during processing then calls next() or responds.

3.1 · Application

2. Salim wants product page 7 via /produits/7. He writes…

app.imprimer('/produits', ...); app.eteindre('/produits', ...); app.get('/produits/:id', (req, res) => res.json({ id: req.params.id })); app.couleur('/produits', ...);

Dynamic route: the segment is captured in req.params.id.

3.1 · Arbitration

3. Salim’s request stays stuck with no response. Correct decision?

Change the terminal colours Check each middleware calls next() or responds, and the parse-before-validate order Reprint everything to reread Delete Express and go back to paper

A forgotten next() blocks the pipeline; validating before parsing reads emptiness.

3.2 · Recall

4. express.static('public') and res.render() respectively…

Delete then recreate the database Switch Node off then on Print then colour the server Serve fixed files, then render an HTML view with data

Static for fixed assets, templates for dynamic HTML (SSR).

3.2 · Application

5. Salim renders the product page in EJS. He writes…

res.render('produit', { produit }); res.imprimer('produit'); app.effacer('produit'); res.couleur('produit');

The views/ view receives the data and produces the HTML.

3.2 · Application

6. Showcase pages to reference + highly interactive dashboard. Salim decides…

SPA everywhere even for three fixed pages Paper for the pages, word of mouth for the rest SSR (EJS) for SEO pages, JSON API + front end for the dashboard No rendering, all in memory

Content/SEO → SSR; interactive app → API + separate front end.

3.3 · Recall

7. MongoDB versus SQL is first of all…

Colour prints in collections JSON/BSON documents in collections, with no rigid schema Network cables in collections No model difference at all

Flexible documents close to JS, versus strict tables and rows.

3.3 · Application

8. Salim enforces a required, unique email on his users. He writes…

email: accept anything with no rule email: deleted on each save email: { type: String, required: true, unique: true } in the schema email: 'roughly required' in a comment

The Mongoose schema validates before saving.

3.3 · Arbitration

9. Salim hesitates between MongoDB and PostgreSQL for linked invoicing. Correct decision?

Neither, all in memory Always MongoDB for everything Always SQL for everything Flexible, evolving data → MongoDB; linked data with critical integrity → SQL

Choose by data nature, not by fashion.

3.4 · Application

10. Salim lists active clients with Prisma. He writes…

await prisma.client.findMany({ where: { actif: true } }); await prisma.imprimer({ tout: true }); await prisma.effacer({ tout: true }); await prisma.colorier({ tout: true });

Typed ORM method generating safe parameterised SQL.

3.4 · Application

11. In raw SQL (pg), Salim filters by entered email. Safe version?

await pool.query('SELECT * FROM clients WHERE email = ' + emailSaisi); await pool.imprimer(emailSaisi); await pool.effacer(emailSaisi); await pool.query('SELECT * FROM clients WHERE email = $1', [emailSaisi]);

The $1 placeholder separates code from data: no injection.

3.4 · Arbitration

12. A colleague suggests concatenating input “just for testing”. Correct decision?

Refuse: parameterisation or ORM always, even in tests Concatenate then print to check Write the real password in the code to go fast Accept, a test is never attacked

Injection is born from concatenation; the safe habit allows no exception.

What's next?

Score ≥ 80%: move on to Level 4 — full REST API, JWT auth and validation (building on sound foundations).
Score 50-79%: revisit the failed sections (pipeline in 3.1, SSR vs API in 3.2, schema in 3.3, parameterisation in 3.4) then retake the test.
Score < 50%: reread level 2 (native http, async/await) then replay each 48h challenge in order.