Choisir entre PostgreSQL et MySQL est l'une des décisions les plus importantes en développement backend.
Vue d'ensemble rapide
PostgreSQL privilégie la conformité aux standards, l'intégrité des données et l'extensibilité. MySQL privilégie la simplicité et la vitesse pour les workloads à forte lecture.
Feature PostgreSQL MySQL 8.x
------------------------------------------------------------
License PostgreSQL (free) GPL / Commercial
ACID Compliance Full Full (InnoDB)
JSON Support JSONB (excellent) JSON (good)
Full-Text Search Built-in tsvector FULLTEXT index
Replication Streaming + Logical Binary log + GTID
Max DB Size Unlimited 256TB
Partitioning Declarative RANGE/LIST/HASH
Window Functions Full support Partial (8.x+)
Extensions Rich ecosystem Plugins (fewer)
Default in Cloud Supabase, RDS PlanetScale, RDSComparaison des fonctionnalités
PostgreSQL a historiquement un ensemble de fonctionnalités beaucoup plus riche.
-- Feature Comparison: PostgreSQL vs MySQL
-- 1. JSON Support
-- PostgreSQL JSONB (binary, indexed)
SELECT data->>'name' FROM users WHERE data @> '{"active": true}';
CREATE INDEX idx_users_data ON users USING GIN (data);
-- MySQL JSON
SELECT JSON_EXTRACT(data, '$.name') FROM users
WHERE JSON_EXTRACT(data, '$.active') = true;
-- 2. Full-Text Search
-- PostgreSQL (built-in tsvector)
SELECT * FROM articles
WHERE to_tsvector('english', content) @@ to_tsquery('postgresql & performance');
-- MySQL FULLTEXT
SELECT * FROM articles
WHERE MATCH(content) AGAINST ('postgresql performance' IN NATURAL LANGUAGE MODE);
-- 3. CTEs (Common Table Expressions)
-- PostgreSQL supports recursive CTEs natively (since 8.4)
WITH RECURSIVE category_tree AS (
SELECT id, name, parent_id, 0 AS level
FROM categories WHERE parent_id IS NULL
UNION ALL
SELECT c.id, c.name, c.parent_id, ct.level + 1
FROM categories c JOIN category_tree ct ON c.parent_id = ct.id
)
SELECT * FROM category_tree ORDER BY level, name;
-- 4. Window Functions (both support, PostgreSQL more complete)
SELECT
employee_id,
salary,
AVG(salary) OVER (PARTITION BY department_id) as dept_avg,
RANK() OVER (PARTITION BY department_id ORDER BY salary DESC) as rank
FROM employees;Comparaison des performances
Les performances brutes dépendent fortement du type de workload.
-- Performance Tuning Examples
-- PostgreSQL: EXPLAIN ANALYZE
EXPLAIN (ANALYZE, BUFFERS, FORMAT JSON)
SELECT u.id, u.name, COUNT(o.id) as order_count
FROM users u
LEFT JOIN orders o ON u.id = o.user_id
WHERE u.created_at > '2025-01-01'
GROUP BY u.id, u.name
ORDER BY order_count DESC
LIMIT 100;
-- PostgreSQL-specific indexes
CREATE INDEX CONCURRENTLY idx_orders_user_status
ON orders (user_id, status)
WHERE status != 'cancelled'; -- Partial index
CREATE INDEX idx_products_attrs ON products USING GIN (attributes); -- JSONB index
-- MySQL: EXPLAIN FORMAT=JSON
EXPLAIN FORMAT=JSON
SELECT u.id, u.name, COUNT(o.id) as order_count
FROM users u
LEFT JOIN orders o ON u.id = o.user_id
WHERE u.created_at > '2025-01-01'
GROUP BY u.id, u.name
ORDER BY order_count DESC
LIMIT 100;
-- Connection pooling configuration
-- PostgreSQL (PgBouncer or pg pool settings)
max_connections = 100 -- in postgresql.conf
-- Use connection pooler for high concurrency
-- PgBouncer pool_mode = transaction (recommended for web apps)
-- MySQL
max_connections = 200 -- in my.cnf
innodb_buffer_pool_size = 4G -- 70-80% of RAM for dedicated MySQL serverQuand choisir PostgreSQL
PostgreSQL est le bon choix quand vous avez besoin de fonctionnalités SQL avancées.
- Complex queries — Advanced window functions, CTEs, lateral joins
- JSON/document storage — JSONB with GIN indexes rivals MongoDB
- Geospatial data — PostGIS extension is the gold standard
- Strict data integrity — CHECK constraints, exclusion constraints, custom domains
- High-concurrency writes — MVCC handles concurrent writes better than MySQL
- Full-text search — Built-in, no external service needed
- Analytics/reporting — Better at complex aggregations and window functions
Quand choisir MySQL
MySQL reste un excellent choix pour les applications web à forte lecture.
- Simple CRUD applications — WordPress, Drupal, and many CMSes default to MySQL
- Read-heavy workloads — InnoDB is highly optimized for read-heavy patterns
- Existing MySQL ecosystem — If your team has deep MySQL expertise
- PlanetScale — MySQL-compatible serverless database with excellent DX
- Replication simplicity — MySQL replication is well-understood and widely deployed
Migration entre les bases de données
Migrer de MySQL à PostgreSQL nécessite une planification soigneuse.
-- MySQL to PostgreSQL Migration: Common Differences
-- 1. AUTO_INCREMENT -> SERIAL/IDENTITY
-- MySQL
CREATE TABLE users (
id INT AUTO_INCREMENT PRIMARY KEY,
name VARCHAR(100)
);
-- PostgreSQL
CREATE TABLE users (
id SERIAL PRIMARY KEY, -- or BIGSERIAL for large tables
name VARCHAR(100)
);
-- Modern PostgreSQL (v10+):
-- id INT GENERATED ALWAYS AS IDENTITY PRIMARY KEY
-- 2. String functions differ
-- MySQL: IFNULL
SELECT IFNULL(phone, 'N/A') FROM users;
-- PostgreSQL: COALESCE (also works in MySQL)
SELECT COALESCE(phone, 'N/A') FROM users;
-- 3. String concatenation
-- MySQL
SELECT CONCAT(first_name, ' ', last_name) FROM users;
-- PostgreSQL
SELECT first_name || ' ' || last_name FROM users;
-- Or use CONCAT (PostgreSQL also supports it)
-- 4. LIMIT/OFFSET syntax (both support the same)
SELECT * FROM users ORDER BY id LIMIT 10 OFFSET 20;
-- 5. Boolean values
-- MySQL: TRUE/FALSE or 1/0
-- PostgreSQL: TRUE/FALSE or 't'/'f' or 'true'/'false'
-- 6. Timestamp with timezone
-- MySQL: DATETIME or TIMESTAMP (local time only)
-- PostgreSQL: TIMESTAMPTZ (timezone-aware, recommended)
ALTER TABLE events
ALTER COLUMN created_at TYPE TIMESTAMPTZ
USING created_at AT TIME ZONE 'UTC';Questions fréquentes
PostgreSQL est-il plus rapide que MySQL ?
Cela dépend du workload. MySQL est souvent plus rapide pour les requêtes simples. PostgreSQL excelle dans les requêtes analytiques complexes.
Quel fournisseur cloud recommande quelle base de données ?
Tous les grands fournisseurs cloud proposent les deux. Supabase est construit sur PostgreSQL, PlanetScale sur MySQL.
PostgreSQL supporte-t-il JSON ?
Oui, PostgreSQL a un excellent support JSON avec le type JSONB, qui stocke JSON en format binaire avec indexation.
Lequel est meilleur pour un nouveau projet en 2026 ?
Pour la plupart des nouveaux projets, PostgreSQL est le choix recommandé par défaut en 2026.