Use pg_dump with the -n flag to export a single PostgreSQL schema to a file for backup or migration.
Exporting a single schema lets you back up or migrate a self-contained part of the database, keeping dump files small and restores fast.
Run pg_dump with the -n option. This flag filters objects so that only the chosen schema is written to the output file.
At minimum supply -U for the role, -h for the host, and -d for the database. PGPASSWORD or a .pgpass file can store the password securely.
Use --schema-only to dump DDL only, --data-only for data only, or omit both to include everything.
Feed the file into psql: psql -U target_user -d target_db -f sales_schema.sql. Ensure the target schema does not already exist, or drop it first.
Compress the dump with -F c (custom format) and enable parallelism with -j. Always test a restore on a staging database before touching production.
Add --blobs to include large objects (BLOBs) in the dump. Without this flag, large objects are skipped.
Create a cron job or CI pipeline that calls pg_dump with the desired flags, pipes the output to gzip, and moves it to durable storage such as S3.
When restoring, explicitly set SET search_path TO your_schema or objects may be created in public.
Always pass the correct -d value. Listing databases with \l in psql beforehand prevents surprises.
Yes. Repeat the -n flag for each schema: -n sales -n inventory.
Only briefly. It takes a SHARE LOCK at the start of the dump to ensure a consistent snapshot, but normal reads and writes continue.
Pipe the output through openssl or gpg: pg_dump ... | gzip | gpg -c -o sales_schema.sql.gpg.
.avif)
.avif){kind=logo}
