Repository pattern implementation that knows nothing (Jon Snow) about the database table a...

Do you have a question? Post it now! No Registration Necessary.  Now with pictures!

Threaded View
I've seen around the Internet and Github, implementations for the design pa
ttern Repository that knows about database table and column names. I was th
ink, if I want to work with the database as a plugin, that I can unplug and
 plug another respecting Open/Closed for the rest of my code, my repository
 should not know about the column names of the database I'm using. So how t
o implement this pattern in a way that it can transform the result from the
 database into a Entity of my domain, without knowing about the database ta
ble and column names?

As my main language is PHP, I saw that in Doctrine\ORM you can easily pass  
different yamls or xmls config files, mapping the column names to property  
names on the Entity, but... I cant be a hostage of a library implementation
, if I want to implement raw PDO, or any other library for my repositories  
that doesn't make this hydration out-of-the-box, my implementation should d
o, so how?

Re: Repository pattern implementation that knows nothing (Jon Snow) about the database table and column names

On 7/8/2014 11:47 AM, wrote:
Quoted text here. Click to load it

No, typically the repository DOES know about the table and column names
in the database; it's purpose is to separate the database definition
from the rest of the program so that only the repository depends on the
database.  Change the database and change the repository, but the rest
of the code remains the same.

Without knowing the database information, the repository would have to
query the database definitions and internally build a description of the
database.  This can easily be done (PhpMyAdmin, for instance, does it).
 However, what this doesn't tell the program is exactly what's in each
column and how tables are related.

For instance, take a simple example (shortened for brevity):

TABLE employee

TABLE department

TABLE employee_department

Now, you want a list of all employees managed by "John Doe".  To a
person, constructing the query from the above definition would be
relatively easy.  But it would be very complex for a program to do.

This could be done, as you found, by passing descriptions of the
database.  But you don't want that.

All I can say is, good luck.  It's something which has been dreamed
about for years, but no one has been able to implement except in very
limited circumstances.

Remove the "x" from my email address
Jerry Stuckle

Re: Repository pattern implementation that knows nothing (Jon Snow) about the database table and column names

On 8.7.2014. 17:47, wrote:
Quoted text here. Click to load it

Yes, it's possible.

Firstly, you need to make a technology-agnostic model (no Doctrine  
annotations, etc.), then repository interface.

Later, you can plug in whatever you want. That plugs are implementations...

Let's say, one implementation could be made using Doctrine 2 ORM.  
Because your model is tech-agnostic, you can't use annotations to map  
it, but nothing stop you to use external XML maps. Beside mapping, you  
need to create a repository implementation using Doctrin 2 ORM. Here is  
the skeleton of such implementation:

class DoctrineUserRepository implements UserRepository
     private $doctrineRepository;

     public function __construct(EntityRepository $doctrineRepository)
         $this->doctrineRepository = $doctrineRepository;

     // ...

Only things you should keep in mind are technology limitations - you  
can't make naive model (without investigating Doctrine ORM limitations)  
because you will encounter those limitations later (inheritance quirks,  
compostite primary keys, value objects...)

Site Timeline