4 ms·
Don't use PostgreSQL or MySQL/MariaDB for analytic work - they're super-slow at that. Try systems like MonetDB, ClickHouse (not a full DBMS) etc: https://en.wik
by einpoklum 7y ago
Don't use PostgreSQL or MySQL/MariaDB for analytic work - they're super-slow at that. Try systems like MonetDB, ClickHouse (not a full DBMS) etc:
https://en.wikipedia.org/wiki/List_of_column-oriented_DBMSes https://en.wikipedia.org/wiki/List_of_column-oriented_DBMSes
Column stores are where it's at.
(Not relevant if you need to process transactions.)
- stingraycharles 7y agoAs someone who works for a high perf timeseries database (QuasarDB), the first thing I tell potential clients is to solve the problem with PostgreSQL. For people who are still exploring their problem domain, using something as flexible as Postgres is a blessing. You’ll be surprised how far you can get nowadays in terms of performance with it. Only once you properly understand your problem domain, and more precisely the limitations of PostgreSQL that you are running into, you can start figuring out your actual requirements and look for an appropriate database that goes with it.