How to Use ParadeDB Enterprise Edition in PostgreSQL

How do I install and use ParadeDB Enterprise Edition in PostgreSQL?

ParadeDB Enterprise Edition is a PostgreSQL extension that adds high-performance vector search, semantic ranking, and advanced full-text capabilities.

Welcome to the Galaxy, Guardian!
You'll be receiving a confirmation email

Follow us on twitter :)

Oops! Something went wrong while submitting the form.

Description

Example H2

Example H3

What problems does ParadeDB Enterprise Edition solve?

ParadeDB EE brings lightning-fast vector search, ML-powered ranking, and incremental indexing directly into PostgreSQL, eliminating the need for a separate search service.

How do I install ParadeDB Enterprise Edition?

Superusers install the shared library, then run CREATE EXTENSION paradedb;. To unlock enterprise features, set paradedb.license_key and reload.

Installation steps

1. Copy paradedb.so to $PGHOME/lib.
2. Add shared_preload_libraries = 'paradedb' in postgresql.conf.
3. ALTER SYSTEM SET paradedb.license_key = 'YOUR-KEY';
4. Restart PostgreSQL.
5.CREATE EXTENSION paradedb;

What is the syntax for creating a searchable vector column?

Create a VECTOR column and an index with ParadeDB’s operator class.

ALTER TABLE products ADD COLUMN embedding VECTOR(1536); CREATE INDEX products_embedding_idx ON products USING paradedb_ivfflat (embedding);

How do I insert vectors?

Convert model output to an array and cast to vector.

INSERT INTO products (id,name,price,stock,embedding) VALUES (42,'Noise-Canceling Headphones',199.99,25, '[0.12,-0.04, ... ,0.88]'::vector);

How do I run a semantic product search?

Use the <=> distance operator in the ORDER BY clause.

SELECT id,name,price FROM products ORDER BY embedding <=> '[0.11,-0.07,...]'::vector LIMIT 10;

Can I combine vector and keyword filters?

Yes—stack ParadeDB distance with normal WHERE clauses.

SELECT id,name FROM products WHERE stock > 0 ORDER BY embedding <=> '[0.11,-0.07,...]'::vector LIMIT 5;

How do I set up hybrid full-text + vector ranking?

Use paradedb.rank() to blend tsvector and vector scores.

SELECT id,name, paradedb.rank(tsv,'camera'::tsquery,embedding, '[0.09,0.02,...]'::vector) AS score FROM products ORDER BY score DESC LIMIT 20;

Best practices for ParadeDB EE?

• Use VECTOR(1536) for OpenAI embeddings.
• Keep index lists between 1–4× rows^0.5 for IVFFLAT.
• Refresh materialized views off-peak.
• Store embeddings in a separate table for large catalogs.

Common pitfalls and fixes

Incorrect dimension: Vectors must match column dimension.Cast with the correct length.
Missing shared_preload_libraries: ParadeDB won’t load; edit postgresql.conf and restart.

Need to re-index after bulk loads?

Run ALTER INDEX ... REBUILD or REINDEX INDEX ... to optimize IVFFLAT clusters after large inserts.

Why How to Use ParadeDB Enterprise Edition in PostgreSQL is important

How to Use ParadeDB Enterprise Edition in PostgreSQL Example Usage


-- Find the 5 most semantically similar products to a customer’s last purchase
WITH last_item AS (
  SELECT p.embedding
  FROM orders o
  JOIN orderitems oi ON oi.order_id = o.id
  JOIN products p     ON p.id = oi.product_id
  WHERE o.customer_id = 17
  ORDER BY o.order_date DESC
  LIMIT 1
)
SELECT id,name,price
FROM products, last_item
ORDER BY embedding <=> last_item.embedding
LIMIT 5;

How to Use ParadeDB Enterprise Edition in PostgreSQL Syntax


-- Enable the extension (superuser)
CREATE EXTENSION paradedb [WITH VERSION 'x.y.z'];

-- Create a vector column on Products
after adding embeddings
ALTER TABLE products ADD COLUMN embedding VECTOR(dimension);

-- Create ParadeDB IVFFLAT or HNSW index
CREATE INDEX name_idx ON products USING paradedb_ivfflat (embedding)
WITH (lists = 100);
-- or
CREATE INDEX name_idx ON products USING paradedb_hnsw (embedding)
WITH (M = 16, ef_construction = 200);

-- Query for nearest neighbors
SELECT id, name, price
FROM products
ORDER BY embedding <=> '[0.1,0.2,...]'::vector
LIMIT 10;

-- Combine with filters
SELECT *
FROM orders o
JOIN customers c ON c.id = o.customer_id
WHERE o.order_date >= CURRENT_DATE - INTERVAL '30 days'
ORDER BY o.total_amount DESC, o.embedding <=> '[...]'::vector
LIMIT 20;

Common Mistakes

Using an embedding vector with the wrong dimension (e.g., 768 into VECTOR(1536)); ParadeDB raises "dimension mismatch". Always ensure your ARRAY length matches the column definition.
Forgetting to set shared_preload_libraries = 'paradedb', causing CREATE EXTENSION to fail. Edit postgresql.conf, add the setting, and restart PostgreSQL before running CREATE EXTENSION.