Vector Search with Relational Databases using PostgreSQL

Overview

With vector databases now powering business competitiveness through super-fast applications such as recommendation engines, it’s no surprise that the vector database market is set to grow 23% CAGR by 2032 (Markets and Markets). This micro course gives aspiring data scientists, ML engineers, gen-AI engineers, software developers, and other data-oriented roles the in-demand skills for performing vector searches in relational databases. Businesses use vector search with relational databases to improve information retrieval via advanced similarity matching. You’ll gain hands-on experience working with PostgreSQL as your relational database platform and Python and JavaScript to vectorize data, create embeddings and collections, and load data, including bulk insertion techniques. Plus, you’ll provide similarity search recommendations using techniques such as cosine similarity. This micro course is part of the IBM Vector Database Fundamentals specialization, designed for professionals building on their NoSQL and relational database experience to work with vector databases. So, enroll today and get set to power your career with highly sought-after relational vector database skills.

Syllabus

Vector Search Practices for SQL Databases

Welcome to this module, where you’ll learn how to implement vector searches using relational databases. You’ll begin with a recap of RDBMS and then dive into the structures RDBMS uses to support vector data types and queries. You’ll apply what you know to perform similarity search tasks using special operators available in PostgreSQL. And, with a focus on PostgreSQL, you’ll learn how to create tsvector data, perform tsquery tasks, and perform bulk inserts using pg-vector for Node.js and psycopg2 for Python.