This tutorial has a new version, check it out!
> Symfony 5 > Go Pro with Doctrine Queries (Legacy)
Login to bookmark this video
Buy Access to Course
04.

JOINs!

Share this awesome video!

|

Keep on Learning!

Start your All-Access Pass

I love JOINs. I do! I mean, a query isn't truly interesting unless you're joining across tables to do some query Kung fu. Doctrine makes JOINs really easy - it's one of my favorite features! Heck, they're so easy that I think it confuses people. Let me show you.

Right now, our search matches fields on the Category, but it doesn't match any of the fortunes in that Category. So if we search for cat, we get the no-results frowny face. Time to fix it!

LEFT JOIN in SQL

The query for this page is built in the search() function. Let's think about what we need in SQL first. That query would select FROM category, but with a LEFT JOIN over to the fortune_cookie table ON fortune_cookie.categoryId = category.id. Once we have the LEFT JOIN in normal SQL land, we can add a WHERE statement to search on any column in the fortune_cookie table.

SELECT cat.* FROM category cat
    LEFT JOIN fortune_cookie fc ON fc.categoryId = cat.id
    WHERE fc.fortune LIKE '%cat%';

ManyToOne and OneToMany Mapping

In Doctrine-entity-land, all the relationships are setup. The FortuneCookie has a ManyToOne relationship on a category property:

176 lines | src/AppBundle/Entity/FortuneCookie.php
// ... lines 1 - 12
class FortuneCookie
{
// ... lines 15 - 23
/**
* @var Category
*
* @ORM\ManyToOne(targetEntity="Category", inversedBy="fortuneCookies")
* @ORM\JoinColumn(nullable=false)
*/
private $category;
// ... lines 31 - 174
}

And inside Category, we have the inverse side of the relationship: a OneToMany on a property called fortuneCookies:

113 lines | src/AppBundle/Entity/Category.php
// ... lines 1 - 13
class Category
{
// ... lines 16 - 38
/**
* @ORM\OneToMany(targetEntity="FortuneCookie", mappedBy="category")
*/
private $fortuneCookies;
// ... lines 43 - 111
}

Mapping this side of the relationship is optional, but we'll need it to do our query.

Adding the leftJoin Query

Let's go add our LEFT JOIN to the query builder! If you're thinking there's a leftJoin method, winner! And this time, we are going to use it. Join on cat.fortuneCookies. Why fortuneCookies? Because this is the name of the property on Category for this relationship.

The second argument to leftJoin() is the alias we want to give to FortuneCookie, fc::

36 lines | src/AppBundle/Entity/CategoryRepository.php
// ... lines 1 - 23
public function search($term)
{
return $this->createQueryBuilder('cat')
// ... lines 27 - 29
->leftJoin('cat.fortuneCookies', 'fc')
// ... lines 31 - 33
}
// ... lines 35 - 36

And right away, we can see why Doctrine JOINs are so easy, I mean confusing, I mean easy. This is all we need for a JOIN - we don't have any of the LEFT JOIN ON fortune_cookie.categoryId = category.id kind of stuff. Sure, this will be in the final query, but we don't need to worry about that stuff because Doctrine already knows how to join across this relationship: all the details it needs are in the relationship's annotations.

The cat.fortuneCookies thing only works because we have the fortuneCookies OneToMany side of the relationship. Adding this mapping for the inverse side is optional, but if we didn't have it, we'd need to add it right now: our query depends on it.

LEFT JOIN, check! And just like normal SQL, we can use the fc alias from the joined table to update the WHERE clause. I'll break this onto multiple lines for my personal sanity and then add OR fc.fortune LIKE :searchTerm because fortune is the name of the property on FortuneCookie that holds the message:

36 lines | src/AppBundle/Entity/CategoryRepository.php
// ... lines 1 - 23
public function search($term)
{
return $this->createQueryBuilder('cat')
->andWhere('cat.name LIKE :searchTerm
OR cat.iconKey LIKE :searchTerm
OR fc.fortune LIKE :searchTerm')
->leftJoin('cat.fortuneCookies', 'fc')
->setParameter('searchTerm', '%'.$term.'%')
->getQuery()
->execute();
}
// ... lines 35 - 36

Moment of truth! We've got a match! Our fortunes are being searched.

JOINing, but Querying for the same Data

Even though we now have a LEFT JOIN, the result of the query is no different: it still returns an array of Category objects. We can and will do some JOINs in the future that actually select data from the joined table. But if all you do is JOIN like we're doing here, it doesn't change the data that's returned.