如何扩展Play2的发展

Question

I'm recently started using Play2 on a project, and read the section on evolutions . And while the example they cite seems fine if my project had 1 table, it seems like it would be very messy if I had 10-20 tables in 1.sql and then changes to them split up over 2.sql , 3.sql and so on.

In Ruby on Rails, Symfony, and others, you define your up/down migrations per entity.

My question is, what is the best way to setup your evolutions in Play2? Should I have all my tables in 1.sql and then make little changes to them over 2.sql and so on? Or is there a way to have a separate .sql file for each table?

Also, are there any examples of large, open source Play2 projects where I could see how it would look?

Answer 1

Actually Play has not possibility to divide evolutions by entities.

IMHO it's rather matter of taste, you can add each entity in single next evolution, anyway only difference will be that counter of evolution will be bigger, I don't think that will help you to keep evolutions cleaner.

Typical workflow is starting from ... good planning. Just create some graph representation of your schema and try to add there as many things as you need. It helps a lot while the project startup and also in next steps of development.

If you are gonna to use Ebean, create all models from your graph and let the plugin to create automatic first evolution file, probably you will save a lot of time on writing evolutions for relations, constraints, etc. Spend some time for fixing and checking initial schema before further development.

After that you need to disable automatic updates as they drops whole DB and recreates tables them from the scratch (there's no diff schema update in Ebean).

It's also matter of taste but I prefer to combine several changes into single evolutions (so again planning...) instead of creating next and next files for every single change ad hoc.