Oracle & Postgresql DBA Blog
TL;DR: The manual work in an Oracle to PostgreSQL migration depends entirely on how the schema was built — not on its size.
This guide maps three common Oracle schema patterns to the specific migration example that covers each one: legacy OLTP, modern OLTP, and data warehouse.
If you are unsure which applies to your database, the diagnostic section below gives you four questions to find out.
Oracle's CO (Customer Orders) schema is the modern replacement for the older OE schema.
It ships with Oracle 19c, it is actively maintained, and it is built the way most real Oracle applications are built today: IDENTITY columns instead of sequence-trigger pairs, JSON stored in BLOB columns, and views that use Oracle-specific SQL functions.
I ran the full migration using ora2pg 25.0 with Oracle 19c as the source and PostgreSQL 18 as the target.
This post covers the five problems that required manual intervention — and why each one will appear in almost every production schema you migrate.
This is the third post in the series.
The HR schema post covered sequence-trigger patterns, %TYPE parameters, and the FK re-apply bug.
The SH schema post covered partitioned tables, bitmap indexes, and materialized views.
CO introduces three new problem categories not seen in either of those.
Continue reading “Oracle to Postgres Migration: CO Schema Issues and Fixes”