Here's a bit of context first:
I am writing a project in PHP using Zend Framework and I have some E/R diagrams from past meetings with the client. The database schema was written for MySQL and now it is time to implement my models in Zend. I am fairly new to Zend so I do not know how to start writing this code nicely.
The first thing I noticed while reading their documentation is that the quick guide used a data gateway pattern, but their Zend_DB reference page did not.
// Data Gateway Sample
// <application/models/UserMapper.php>
class Application_Model_UserMapper()
{
// ... Methods to read/write from DB
}
// <application/models/DbTable/Users.php>
class Application_Model_DbTable_Users
{
// Relationship infromation etc.
}
// <application/models/User.php>
class Application_Model_User
{
// At last, model specific data
// using UserMapper->find($id, $model);
}
After browsing stackoverflow for a little while in hope of finding pointers in how to most efficiently organize the models, I came across yet another different recommended solution.
Now the solution linked above looks very clean, but it still leaves a single question in my mind before I set off to write my code:
Should I use Row_Abstract subclasses to store the data model, or should I make separate models that have no other purpose but store data retrieved?
The latter one seems like a lot of duplicate effort (Model_DbTable, Model, Mapper, [Row?])
// What I have in mind
// <application/models/DbTable/Users.php>
class Application_Model_DbTable_Users
{
// Relationship infromation etc.
}
// <application/models/User.php
class Application_Model_User extends Zend_Db_Table_Row_Abstract
{
// Model stuff here
// Relationship fetch helpers here
// ...?
}
// Perhaps a Rowset_Abstract class if it is ever needed?
