Hi everyone
I have a question pertaining to the appropriate use cases for packages like DBIx::Class::Schema and friends.
I'm building a reporting framework, using Catalyst as a display/admin interface. The reports themselves are all being stored in a Postgres database. The database is subdivided into schemas, and makes liberal use of views, functions and privilege separation (I'm even using GSSAPI cred. delegation through Apache/Catalyst to Postgres)
Reading through the Catalyst docs (and the book), DBIx::Class and friends are plugged frequently as the most sane way to do the model part of MVC. Upon using DBIx::Class::Schema::Loader, I realized it could not automatically create a model because of my use of schemas in my database.
I've considered my projected workload, which is read often/write infrequently/update almost never. I've also considered the nature of some of my in-progress reports (they leverage functionality in the database like temp tables to generate reports)
Given that I will have to make all my DBIx::Class::Schema objects by hand, will I come out ahead by making a complete model in that manner? Or am I better off sticking with posing relevant queries directly to DBI, and forgoing the use of DBIx::Class?
If anyone has any experience with the +/- of building model classes for this style of database, I'd appreciate your insights.