How to TRUNCATE TABLE in PostgreSQL

What does TRUNCATE TABLE do in PostgreSQL?

TRUNCATE TABLE deletes every row in a table in a single, metadata-level operation. Unlike DELETE, it bypasses row-by-row logging, giving near-instant execution even on millions of records.

How do I quickly empty a table without logging each row?

Use TRUNCATE TABLE when you need a fresh table state and do not require per-row triggers, foreign-key checks, or VACUUM reclamation. It is ideal for nightly ETL staging tables.

What is the basic TRUNCATE TABLE syntax?

Place the TRUNCATE keyword before one or more table names. Add RESTART IDENTITY to reset SERIAL/BIGSERIAL sequences, or CONTINUE IDENTITY to keep them.

Can I truncate multiple tables at once?

Yes. List several tables separated by commas, or use CASCADE to automatically truncate tables that reference them through foreign keys.

How do I reset identity columns after truncation?

Append RESTART IDENTITY so sequences behind SERIAL or IDENTITY columns restart from their start value. Useful for dev and test datasets.

When should I use CASCADE with TRUNCATE?

Use CASCADE only when you are certain that child rows can be lost. For example, truncating Orders might need CASCADE to clear OrderItems.

Best practices for using TRUNCATE TABLE

Wrap TRUNCATE in a transaction so it can be rolled back. Grant the command sparingly; it requires table owner or superuser privileges.

Common pitfalls and how to avoid them

Avoid truncating tables with critical audit data. Remember that triggers do not fire, so downstream logic will not execute.