Why I switched from MongoDB to PostgreSQL (Rasmus Porsager)

Episode Overview

• The speaker recounts his early programming journey, initially choosing the newer MongoDB over SQL, which he considered an "ancient" technology.
• He discusses his "aha" moment: while using the Mongoose library, he realized he was building schemas on top of a schemaless database, thereby nullifying its core benefit.
• The speaker shares a negative experience where a system he built on MongoDB lost its structural integrity after being handed off to a new team.
• This experience, combined with the growing popularity of Postgres around 2016, led him to switch, and he has been a strong advocate for it ever since.

Key Concepts

• SQL vs. NoSQL: A practical comparison based on a developer's long-term experience, contrasting the initial ease of use of MongoDB with the long-term structural benefits of PostgreSQL.
• Mongoose Library: An Object Data Modeling (ODM) library for MongoDB that allows developers to enforce schemas at the application level on top of the schemaless database.
• Schemaless Databases: The core concept of document-based databases like MongoDB, which offer flexibility by not requiring a predefined data structure.
• Data Integrity and Maintainability: The challenge of maintaining a consistent and predictable data structure in a flexible environment over time, especially as a project is handed off between developers.
• Technology Perception: How developer perceptions of "old" versus "new" technologies can influence architectural decisions, sometimes overlooking the proven benefits of established tools like SQL.

Quotes

• At 00:13 - "I thought it was this ancient thing that wasn't a good idea to use anymore." - The speaker explains his initial perception of SQL when he was a new programmer, favoring the novelty of MongoDB.
• At 00:46 - "So the entire idea behind Mongo being schemaless and document-based, I wasn't using it at all for anything." - He describes his realization that using the Mongoose library to create schemas meant he was effectively fighting against, rather than leveraging, MongoDB's core design philosophy.
• At 02:14 - "Why haven't I given this a shot before?" - His thought after finally exploring PostgreSQL and discovering the directness and power of using SQL for creating structured tables, which he found more suitable for his needs.

Takeaways

• Be mindful of defeating the purpose of your tools; if you find yourself enforcing rigid schemas on a schemaless database, consider whether a relational database like Postgres is a more appropriate choice from the start.
• Prioritize long-term maintainability over initial setup speed. A database with enforced structural integrity can prevent data chaos and be easier for future teams to understand and build upon.
• Re-evaluate established technologies periodically. "Old" technologies like SQL have remained relevant for decades because they solve a fundamental set of problems effectively and reliably.