Summary
[EuroPython 2012] H Krossing - 4 JULY 2012 in "Track Tagliatelle"
Description
Presenting a better way of using PostgreSQL from python with Simple to use way to avoid moving data back and forth to client for just for processing while still maintaining your code in client code and getting all the benefits of server-side programming with ease of Python. access to data without writing SQL for simple usage while SQL is still available where appropriate, like complex data analysis This is achieved by the following: decorators which ship python code to be executed in database, with potentially 100x speedup for some use cases direct building of query plans bypassing SQL generation and parsing Some more infrastructure is presented which is needed for using code-shipping in more complex scenarios like django extra import hooks in server for loading python modules from caller (client) extra import hooks for importing modules from database loadatstartup modules for pl/python ways to scale this type of application by sharding (a.k.a. horisontally scaling) it using pl/proxy