February 2021 Summaries

This article provides a comprehensive guide to prepare for PostgreSQL interviews, focusing on transactional SQL aspects. It covers essential topics such as modeling relationships, using rich type systems, namespacing with SCHEMAs, indexing strategies, creating VIEWs, managing roles and permissions, implementing row-level security policies, writing functions in PL/pgSQL, thinking in SETs, and utilizing common table expressions (CTEs). The article emphasizes the importance of practical experience and suggests using Supabase for hands-on practice.

This comprehensive guide delves into key PostgreSQL topics that are essential for acing technical interviews, emphasizing the transactional aspects of the database system. It highlights the importance of understanding data modeling, including different relationship types and the use of foreign keys, as well as the utilization of PostgreSQL's rich type system like arrays and JSONB. The guide also covers the significance of indexing for optimizing query performance and the use of EXPLAIN ANALYZE for query analysis. It discusses the creation and management of views for data representation, including the concept of autoupdatable views and their limitations. Security aspects such as roles, permissions, and row-level security (RLS) are also explored, along with the use of functions and PL/pgSQL for implementing business logic. Additionally, the guide emphasizes thinking in sets and the benefits of using common table expressions (CTEs) for more readable queries. It encourages practical application of these concepts to truly master PostgreSQL, with a nod to Supabase as a platform for quickly setting up and experimenting with PostgreSQL databases.

Brad Dwyer, founder of Roboflow, a startup that helps developers integrate computer vision into their applications, used Supabase to launch Paint.wtf, a product that survived traffic from the front page of Hacker News, Reddit, and Product Hunt. The team built Paint.wtf in a weekend using OpenAI's new CLIP model, which classifies images by flipping image classification into a text similarity task. Roboflow chose Supabase for its leaderboard feature due to its built-in functionality for counting the number of documents in a collection. The product received significant traction and sustained coverage in the media, with over 100K users submitting drawings within a 24-hour span.

Brad Dwyer, founder of Roboflow, a startup that helps developers integrate computer vision into applications, used Supabase to build Paint.wtf, a product that survived traffic from the front pages of Hacker News, Reddit, and Product Hunt. The team leveraged OpenAI's new CLIP model for their weekend project, which classifies images by converting image classification into a text similarity task. They built a leaderboard using Supabase to count documents in collections, enabling users to track their performance against others. Paint.wtf received significant attention and sustained coverage, with over 100K users submitting drawings within 24 hours.

In January, Supabase introduced several new features and improvements to its platform. These include count functionality support for PostgreSQL, two new Auth providers (Facebook and Azure), an audit trail exposed in the dashboard, a React Auth Widget, a new auth.email() function, expanded regions for database launching, and the ability to copy SQL results as Markdown. Additionally, Supabase has made strides in community development with several open-source libraries and projects. The company is preparing for "Launch week" at the end of Q1, which will introduce more features such as Storage.

In January, Supabase introduced several new features and improvements. These include count functionality for PostgreSQL, support for Facebook and Azure Auth providers, an audit trail in the dashboard, a React Auth UI widget, a new auth.email() function, and the ability to launch databases in London or Sydney. Additionally, SQL results can now be copied as Markdown, and Supabase supports React Server components. The community has also made significant advances on various libraries. Upcoming features include Storage during "Launch week" at the end of Q1.