Question 1

What is PL/Python?

Accepted Answer

PL/Python is a PostgreSQL extension that lets you write functions and stored procedures in Python — directly inside the database engine. Two variants ship: plpython3u (untrusted, full OS access — superuser only) and historically plpython3 (trusted, sandboxed — deprecated since PG 13). You declare a function with CREATE FUNCTION ... LANGUAGE plpython3u AS $$ ... $$;, then call it from SQL like any other function. Common use cases: data cleaning with Pandas, ML inference, calling external libraries that lack Postgres-native equivalents.

Question 2

What's the difference between PL/pgSQL and PL/Python?

Accepted Answer

PL/pgSQL is Postgres' native procedural language — SQL with control flow (loops, conditionals, exception handling). Fast, type-safe, integrates naturally with SQL, but limited to what Postgres can express directly. PL/Python lets you write functions in Python — slower (~10-100x for tight loops) but with access to the entire Python ecosystem. Rule of thumb: use PL/pgSQL for business logic that's mostly SQL with some control flow. Use PL/Python when you need libraries (Pandas, NumPy, scikit-learn) or string-processing capabilities SQL lacks.

Question 3

Can I use JavaScript in Postgres?

Accepted Answer

Yes — install the plv8 extension (V8-powered JavaScript). It's especially popular for JSON-heavy workloads since V8 handles JSON natively and faster than PL/pgSQL. Common uses: writing JSON transformation functions, sharing validation logic between frontend and database, and complex JSON path operations. plv8 supports modern JS (ES2020+) and can use NPM-style modules with some setup. Supabase ships plv8 enabled by default.

Question 4

Does PL/Python work on AWS RDS?

Accepted Answer

Yes — AWS RDS supports plpython3u on PostgreSQL 12+. Enable it by adding plpython3u to the database via CREATE EXTENSION plpython3u; from a superuser-equivalent role. Aurora Postgres supports it similarly. Note that plpython3u is the untrusted variant (the only one available on modern Postgres) and requires superuser privileges to create functions — managed providers handle this by giving you the rds_superuser role. Supabase and Neon also support plpython3u on most plans.

Question 5

When should I use a stored procedure vs application code?

Accepted Answer

Use stored procedures when: logic is data-intensive and pushing it into the database avoids round-trips (e.g. complex multi-table updates in one transaction), the logic is enforced as an integrity rule that must run regardless of the caller (triggers, RLS policies), or you need to share logic across multiple application services hitting the same database. Use application code when: logic changes frequently, you want easy version control and CI/CD, you need observability tools and APM integration, or you might switch databases later. Most teams should prefer application code as the default and reach for stored procedures only for the specific cases above.

42+ Procedural Languages Extensions for PostgreSQL

What is a PostgreSQL Procedural Language Extension?

When to Add a Procedural Language Extension

Frequently Asked Questions

Manage PostgreSQL Visually