Snowflake Review

Snowflake is my personal favorite data warehouse. They were one of the first to offer a scalable service that separates storage from compute. This architecture is outlined in a whitepaper that I recommend every newbie data engineer reads.

Snowflake was not the first piece of software to read and write columnar files to object storage for fast analytical queries - Hive and Bigquery both predate it by a few years. However, in my opinion, their implementation of the concept is excellent and its nature as a multi-cloud service helped it grab an early lead in the hearts and minds of developers.

Snowflake is frequently compared to Databricks. However, I think their products complement each other, and aren't really direct competitors.

Pros

• It is a modern data warehouse that offers great performance on analytical queries, and separates storage from compute.
• Multi cloud! It is available on AWS, Azure, and GCP.
• They have business intelligence dashboards built right in to the web UI.
• The Snowflake documentation is among the very best I've ever seen. Right up there with PostgreSQL.
• Great sales and customer service team.
• The whole ecosystem's UX / UI is quite pleasant. Kudos to their designers.

Cons

• Snowflake doesn't offer an ELT tool. Honestly, this bewilders me. I've built a simple ELT tool myself, so I know it isn't that hard. They definitely have the talent and resources to build one.
• Its pricing is credit based. No one ever really understands what a credit is, other than Snowflake salespeople. I know that. You know that. Snowflake knows that. Tech companies can take their credits and get lost, as far as I'm concerned.