Should I choose a columnar or relational database for analytics?

[Charlesjone]

There are many factors to consider between a columnar database vs relational database and it will depend more heavily on how the workload will be used. Typically, relational systems work very well in transactional applications, where each record is accessed and altered frequently. In general, column-based systems are well-suited for analytics queries, which process and aggregate a lot of data. A column-oriented approach has the potential for considerable performance benefits for reporting and business intelligence. But there is no clear victor in the end. When selecting an architecture, Datalance can assess various factors, like query patterns, scalability, data size, and current infrastructure.

[classshot]

Solid points. Nowadays, many teams also look into HTAP solutions or using logical replication (CDC) to pipe transactional data from an RDBMS into a columnar data warehouse. That way, you avoid overloading production OLTP systems while getting the best query performance for BI and reporting.