Fiche mémo PostgreSQL : guide de référence rapide pour les développeurs

# Connexion
psql -h HOST -p 5432 -U USER -d DB
psql $DATABASE_URL

# Dans psql
\conninfo             -- afficher la connexion
\l[+]                 -- lister les bases de données
\c DB                 -- se connecter à la base DB
\dt[+] [schema.]pat   -- lister les tables
\dv[+]                -- lister les vues
\ds[+]                -- lister les séquences
\df[+] [pat]          -- lister les fonctions
\d[S+] name           -- décrire table/vue/séquence
\dn[+]                -- lister les schémas
\du[+]                -- lister les rôles
\timing               -- activer/désactiver le temps de requête
\x                    -- affichage étendu
\e                    -- éditer le buffer dans $EDITOR
\i file.sql           -- exécuter le fichier
\copy ...             -- COPY côté client
\! shell_cmd          -- exécuter une commande shell

SELECT version();

pg_config --version

psql --version

-- Création du schéma et de la table
CREATE SCHEMA IF NOT EXISTS app;
CREATE TABLE app.users (
  id           bigserial PRIMARY KEY,
  email        text NOT NULL UNIQUE,
  name         text,
  active       boolean NOT NULL DEFAULT true,
  created_at   timestamptz NOT NULL DEFAULT now(),
  profile      jsonb,
  tags         text[]
);

-- Modification
ALTER TABLE app.users ADD COLUMN last_login timestamptz;
ALTER TABLE app.users ALTER COLUMN name SET NOT NULL;
ALTER TABLE app.users DROP COLUMN tags;

-- Contraintes
ALTER TABLE app.users ADD CONSTRAINT email_lower_uk UNIQUE (lower(email));

-- Index
CREATE INDEX ON app.users (email);
CREATE UNIQUE INDEX CONCURRENTLY users_email_uidx ON app.users (lower(email));
CREATE INDEX users_profile_gin ON app.users USING gin (profile);
CREATE INDEX users_created_at_idx ON app.users (created_at DESC);

INSERT INTO app.users (email, name) VALUES
  ('a@x.com','A'), ('b@x.com','B')
RETURNING id;

-- Upsert (ON CONFLICT)
INSERT INTO app.users (email, name)
VALUES ('a@x.com','Alice')
ON CONFLICT (email)
DO UPDATE SET name = EXCLUDED.name, updated_at = now();

UPDATE app.users SET active = false WHERE last_login < now() - interval '1 year';
DELETE FROM app.users WHERE active = false AND last_login IS NULL;

SELECT * FROM app.users ORDER BY created_at DESC LIMIT 20 OFFSET 40; -- pagination

-- Filtrage
SELECT * FROM app.users WHERE email ILIKE '%@example.%' AND active;

-- Agrégats & GROUP BY
SELECT active, count(*) AS n
FROM app.users
GROUP BY active
HAVING count(*) > 10;

-- Jointures
SELECT o.id, u.email, o.total
FROM app.orders o
JOIN app.users u ON u.id = o.user_id
LEFT JOIN app.discounts d ON d.id = o.discount_id;

-- DISTINCT ON (spécifique à Postgres)
SELECT DISTINCT ON (user_id) user_id, status, created_at
FROM app.events
ORDER BY user_id, created_at DESC;

-- CTE (Common Table Expressions)
WITH recent AS (
  SELECT * FROM app.orders WHERE created_at > now() - interval '30 days'
)
SELECT count(*) FROM recent;

-- CTE récursif
WITH RECURSIVE t(n) AS (
  SELECT 1
  UNION ALL
  SELECT n+1 FROM t WHERE n < 10
)
SELECT sum(n) FROM t;

SELECT
  user_id,
  created_at,
  sum(total)      OVER (PARTITION BY user_id ORDER BY created_at) AS running_total,
  row_number()    OVER (PARTITION BY user_id ORDER BY created_at) AS rn,
  lag(total, 1)   OVER (PARTITION BY user_id ORDER BY created_at) AS prev_total
FROM app.orders;

-- Extraction
SELECT profile->>'company' AS company FROM app.users;
SELECT profile->'address'->>'city' FROM app.users;

-- Index pour les requêtes jsonb
CREATE INDEX users_profile_company_gin ON app.users USING gin ((profile->>'company'));

-- Existence / Contenance
SELECT * FROM app.users WHERE profile ? 'company';              -- la clé existe
SELECT * FROM app.users WHERE profile @> '{"role":"admin"}'; -- contient

-- Mise à jour jsonb
UPDATE app.users
SET profile = jsonb_set(COALESCE(profile,'{}'::jsonb), '{prefs,theme}', '"dark"', true);

-- appartenance et contenance
SELECT * FROM app.users WHERE 'vip' = ANY(tags);
SELECT * FROM app.users WHERE tags @> ARRAY['beta'];

-- ajout
UPDATE app.users SET tags = array_distinct(tags || ARRAY['vip']);

SELECT now() AT TIME ZONE 'Australia/Melbourne';
SELECT date_trunc('day', created_at) AS d, count(*)
FROM app.users GROUP BY d ORDER BY d;

-- Intervalles
SELECT now() - interval '7 days';

BEGIN;
UPDATE app.accounts SET balance = balance - 100 WHERE id = 1;
UPDATE app.accounts SET balance = balance + 100 WHERE id = 2;
COMMIT; -- ou ROLLBACK

-- Niveau d'isolation
SET TRANSACTION ISOLATION LEVEL REPEATABLE READ;

-- Vérifier les verrous
SELECT * FROM pg_locks l JOIN pg_stat_activity a USING (pid);

-- Création de rôle/utilisateur
CREATE ROLE app_user LOGIN PASSWORD 'secret';
GRANT USAGE ON SCHEMA app TO app_user;
GRANT SELECT, INSERT, UPDATE, DELETE ON ALL TABLES IN SCHEMA app TO app_user;
ALTER DEFAULT PRIVILEGES IN SCHEMA app GRANT SELECT, INSERT, UPDATE, DELETE ON TABLES TO app_user;

-- Côté serveur (nécessite superutilisateur ou permissions appropriées)
COPY app.users TO '/tmp/users.csv' CSV HEADER;
COPY app.users(email,name) FROM '/tmp/users.csv' CSV HEADER;

-- Côté client (psql)
\copy app.users TO 'users.csv' CSV HEADER
\copy app.users(email,name) FROM 'users.csv' CSV HEADER

EXPLAIN (ANALYZE, BUFFERS, VERBOSE) SELECT ...;  -- temps d'exécution réel

-- Vues de statistiques
SELECT * FROM pg_stat_user_tables;
SELECT * FROM pg_stat_statements ORDER BY total_time DESC LIMIT 20; -- nécessite l'extension

-- Maintenance
VACUUM [FULL] [VERBOSE] table_name;
ANALYZE table_name;
REINDEX TABLE table_name;

CREATE EXTENSION IF NOT EXISTS pg_stat_statements;
CREATE EXTENSION IF NOT EXISTS btree_gin;
CREATE EXTENSION IF NOT EXISTS pg_trgm; -- recherche de trigrammes

ALTER TABLE app.docs ADD COLUMN tsv tsvector;
UPDATE app.docs SET tsv = to_tsvector('english', coalesce(title,'') || ' ' || coalesce(body,''));
CREATE INDEX docs_tsv_idx ON app.docs USING gin(tsv);
SELECT id FROM app.docs WHERE tsv @@ plainto_tsquery('english', 'quick brown fox');

\pset pager off       -- désactiver le pager
\pset null '∅'
\pset format aligned  -- autre : unaligned, csv
\set ON_ERROR_STOP on
\timing on

-- Taille des tables
SELECT relname, pg_size_pretty(pg_total_relation_size(relid)) AS size
FROM pg_catalog.pg_statio_user_tables ORDER BY pg_total_relation_size(relid) DESC;

-- Enflure des index (approximatif)
SELECT schemaname, relname, indexrelname, idx_scan
FROM pg_stat_user_indexes ORDER BY idx_scan ASC NULLS FIRST LIMIT 20;

# Dump logique
pg_dump -h HOST -U USER -d DB -F c -f db.dump        # format personnalisé
pg_restore -h HOST -U USER -d NEW_DB -j 4 db.dump

# SQL brut
pg_dump -h HOST -U USER -d DB > dump.sql
psql -h HOST -U USER -d DB -f dump.sql

-- Identité générée
CREATE TABLE t (id bigINT GENERATED ALWAYS AS IDENTITY, ...);

-- UPSERT avec index partiel
CREATE UNIQUE INDEX ON t (key) WHERE is_active;