PostgreSQL is eating the database world

PostgreSQL is not a simple relational database, but an abstract framework for data management with the power to devour the entire database world. And this is also what is happening — “Everything in Postgres” is no longer the cutting-edge exploration of a few elite teams, but has become a best practice that has entered the mainstream view.

OLAP field welcomes kickers

At a database salon in 2016, I made a point: one of the main regrets of the PostgreSQL ecosystem now is the lack of a good enough columnar storage analysis plugin for OLAP analysis. Although PostgreSQL itself provides a powerful set of analysis functions, it is more than enough to cope with routine analysis tasks. However, the performance of full-scale analysis under large data volumes is still somewhat inadequate compared to dedicated real-time data warehouses.

Taking Clickbench, the authoritative evaluation in the field of analysis, as an example, we have marked the performance of PostgreSQL with ecological extension plugins and compatible derivative databases. The native untuned PostgreSQL performs poorly ( x1050 ), but can reach ( x47 ) after tuning; in addition, there are three extensions related to analysis: Hydra ( x42 ), TimescaleDB ( x103 ), and distributed extension Citus ( x262 ).

This kind of analysis performance can’t be said to be bad, because compared with MySQL, MariaDB such as pure OLTP database’s eye-catching performance ( x3065, x19700 ) is indeed much better; but the third-tier…