Switching Teams: Moving an Application from MySQL to PostgreSQL
summary
Migration converts a 2005 MySQL-backed Catme app to PostgreSQL using mysqldump (ANSI mode), sed/perl/C scripts and iconv to repair mixed encodings; enums, AUTO_INCREMENT→sequences, indexes and timestamp behavior are rewritten and many MySQL-specific queries are refactored to standard SQL. Tests run about threefold faster immediately; many tests fail until SQL idioms and date logic are fixed. Production data reveals TIS-620 and CP949 rows that require per-country extraction and recoding with iconv + C fixer; no PostgreSQL extensions are required for full functionality.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 1:15Why switching databases is not straightforward
- 4:02Choosing PostgreSQL over MySQL
- 9:20Schema migration challenges
- 12:15Schema and data type differences
- 16:38Performance improvements
- 17:46Code changes for PostgreSQL
- 22:00Testing production data
- 27:24Fixing non-UTF8 data
- 31:08Q&A
related talks
57mCase Study, Converting the CATME application from MySQL to Postgres
59mLarge Scale MySQL Migration to PostgreSQL
45mSQL Server to PostgreSQL: Migration Challenges & Solutions
27mMigrating from MSSQL to PostgreSQL
48mMigrating to PostgreSQL: Strategies, Sovereignty, and Success
06mMigration Consideration Oracle to PostgreSQL
29mTales from the Field - Oracle to PostgreSQL migrations
43mMigrating Away from MySQL for an awesome full text search
