PGDay Lowlands2026

While the Third Normal Form (3NF) is the gold standard for transactional integrity, it often struggles with complex analytical tasks. Traditionally, bridging the gap between a transactional OLTP model and an analytical Star Schema required the use of third party ETL tools. But as Postgres has matured, the "database-as-a-platform" approach allows us to handle this transformation natively. This talk explores three distinct architectural paths to transforming data from transactional to analytical models within Postgres: 1) Views 2) Materialized Views 3) Triggers All three approaches benefit from a combination with logical replication to build a dedicated analytical replica without impacting transactional performance. The talk dives into the implementation details of each, comparing them across three axes: data freshness, write-amplification cost, and architectural complexity. Attendees will leave with a clear framework for deciding which Postgres technique to use when.