Accelerating PL/pgSQL Code Conversion
summary
Migration often spends roughly 50% of effort on code conversion, especially for projects approaching 1,000,000 lines of code. PL/pgSQL compiles with minimal semantic checks so many errors only surface on first execution (conversion tools report e.g. 90% converted while first-run correctness can be ~20%). Practical accelerators: standardize with wrapper libraries like ORAFCE, run static checks with plpgsql_check, profile nested calls via pg_stat_statements, use Postgres 14’s atomic-function metadata and session warning flags, and enforce strict data-type mapping to avoid expensive rework; cloud managed services may block some extensions.
AI-written summary · about
chapters (6)
- 0:00Introduction to migration
- 10:06Conversion planning essentials
- 20:04Data type mapping strategies
- 27:05Standardizing PL/pgSQL functions
- 32:04Definition of done for conversion
- 38:14Q&A on migration tools
related talks
22mAccelerating PL/pgSQL Code Conversation When Migrating to Postgres
31mSimplifying the Top Complexities While Migrating From Oracle to PostgreSQL
48mMigrating to PostgreSQL: Strategies, Sovereignty, and Success
30mAn Assessment-to-Adoption Path from Oracle, SQL Server, Sybase, DB2 onto
44mDebugging your PL/pgSQL code
24mMigrating from Oracle to Azure Database for PostgreSQL, Seamlessly
01h 09mReplacing Oracle Features in PostgreSQL: HA, Functions & Extensibility
46mThe Top 10 Mistakes When Migrating Oracle Databases to PostgreSQL