Multi-level Partitioning in PostgreSQL
Presented by:
Corey Huinker
Corey Huinker is a database programmer and consultant based in New York City. He specializes in database query optimization and ETLs.
He is the author of the PGXN modules Poor Mans Parallel Processing and range_partitioning.
His past hobbies have included improvisational theater and refereeing roller derby.
No video of the event yet, sorry!
Many RDBMS products offer table partitioning to some limited degree. The number of layers of partitions may be limited to one or two columns, and there may be limited options for partitioning strategy in the second layer (i.e. hash only). PostgreSQL's roll-your-own partitioning creates considerable work for the programmer, but also allows for flexibility in that strategy: asymmetrical partitions, asymmetrical indexing, choices of partitioning strategy, etc. This presentation will retrace the steps taken to find the most suitable methods of implementing partitioning on the data set, how that strategy was implemented, as well as comparing how the partitioning might have looked if it had been done in Oracle or MySQL.
- Date:
- Duration:
- 30 min
- Room:
- Conference:
- PGConf NYC 2014 [PgConf.US]
- Language:
- Track:
- Admin
- Difficulty:
- Medium