Accelerating PL/pgSQL Code Conversation When Migrating to Postgres
summary
Code conversion consumes about 50% of migration effort and the talk lays out concrete accelerators: use plpgsql_check to surface runtime errors before execution, run unit tests with pgTAP and automate type mapping with ora2pg. Plan by dependency graphs and convert low-dependency objects first, mock callable tests from day one, handle empty string vs NULL with NULLIF, pick function argument types deliberately to avoid cast errors, and use plprofiler and pg_stat_statements to find slow nested SQL.
AI-written summary · about
chapters (7)
- 0:00Beginning of Deepak's talk
- 1:01Distribution of Database Migration Efforts
- 2:34Why PL/pgSQL
- 7:12Code conversion needs planning
- 11:32Empty string vs. Null
- 13:51Mapping data types
- 17:31Extensions that help with code conversion
related talks
39mAccelerating PL/pgSQL Code Conversion
48mMigrating to PostgreSQL: Strategies, Sovereignty, and Success
06mMigration Consideration Oracle to PostgreSQL
31mSimplifying the Top Complexities While Migrating From Oracle to PostgreSQL
46mThe Top 10 Mistakes When Migrating Oracle Databases to PostgreSQL
45mBeing mindful of query optimizer
24mMigrating from Oracle to Azure Database for PostgreSQL, Seamlessly
50mHacking on Postgres: An Overview