3 ms·
How were previous implementations of materialized views deficient?
by adamnemecek 6y ago
How were previous implementations of materialized views deficient?
- georgewfraser 6y agoJoins were unavailable or subject to extreme limitations. Or just plain wrong!
- ahupp 6y agoHere's a nice writeup of Materialize: https://lucperkins.dev/blog/new-db-tech-1/#materialize https://lucperkins.dev/blog/new-db-tech-1/#materialize Not really mentioned here, but in standard postgres it might be quite expensive to update the view so you can only do it periodically. Materialize keeps that up-to-date continuously.
- ako 6y agoRDBMSes enable you to create materialized views only for data in the database. Materialize enables you to do this for any streaming data source in your organization, with the ease of writing SQL. This enables you to simply write a SQL statement joining data from Salesforce + SAP + Siebel as soon as the data changes, and store the results as a near real-time up to date database table. It does depend on a lot of underlying plumbing: streaming platform (e.g. kafka), and streaming data sources (e.g., kafka connect + debezium).