[DDC-3767] orm-alternative connection to use different database is working fine but then I have to use orm-default as a driver and not orm-alternative as driver while it should have been the other way. Created: 12/Jun/15  Updated: 12/Jun/15

Status: Open
Project: Doctrine 2 - ORM
Component/s: Mapping Drivers, ORM
Affects Version/s: None
Fix Version/s: None

Type: Bug Priority: Major
Reporter: Ovi Mughal Assignee: Marco Pivetta
Resolution: Unresolved Votes: 0
Labels: mapping
Environment:

Windows 7, Doctrine-ORM-Module 0.9, ZF2 2.5



 Description   

I needed to create module level database connection, I did override to orm-default and created orm-alternative connection but when using driver for it as orm-alternative, object manager instance is not created but when I use orm-default as driver everything works fine. But why?

Here is what is my config:
NOTE: if change orm_oz_default to orm-default on line #25, everything works fine.

return array(
1. 'doctrine' => array(
2. 'connection' => array(
3. 'orm_oz_default' => array(
4. 'driverClass' => 'Doctrine\DBAL\Driver\PDOMySql\Driver',
5. 'params' => array(
6. 'host' => 'localhost',
7. 'port' => '3306',
8. 'user' => 'root',
9. 'password' => '',
10. 'dbname' => 'zfdocttwo',
11. )
12. )
13. ),
14. 'entitymanager' => array(
15. 'orm_oz_default' => array(
16. 'connection' => 'orm_oz_default',
17. )
18. ),
19. 'driver' => array(
20. 'moduletwo_Entity' => array(
21. 'class' => 'Doctrine\ORM\Mapping\Driver\AnnotationDriver',
22. 'cache' => 'array',
23. 'paths' => array(_DIR_.'/../src/Moduletwo/Entity')
24. ),
25. 'orm_oz_default' => array(
26. 'drivers' => array(
27. 'Moduletwo\Entity' => 'moduletwo_Entity'
28. )
28. )
30. )
31. ),
32.);






[DDC-3735] Problem with Collate Created: 13/May/15  Updated: 13/May/15

Status: Open
Project: Doctrine 2 - ORM
Component/s: Mapping Drivers, ORM
Affects Version/s: 2.5
Fix Version/s: None

Type: Bug Priority: Major
Reporter: Hugo Henrique Assignee: Marco Pivetta
Resolution: Unresolved Votes: 0
Labels: dbal, mapping, mysql, schematool
Environment:

development



 Description   

I'm using Migrations and always when a new version changes in my schema in action `up` SQL removes the definition of the table COLLATE for example:

Version20150513194922.php
public function up(Schema $schema)
{
    $this->addSql('ALTER TABLE customer CHANGE user_id user_id CHAR(36) DEFAULT NULL COMMENT \'(DC2Type: guid)\'');
}

public function down(Schema $schema)
{
    $this->addSql('ALTER TABLE customer CHANGE user_id user_id CHAR(36) NULL DEFAULT COLLATE utf8_unicode_ci COMMENT \'(DC2Type: guid)\'');
}
User.php
/**
 * @ORM\Table(name="user")
 * @ORM\HasLifecycleCallbacks
 */
class User
{
    /**
     * @ORM\Id
     * @ORM\Column(type="guid", options={"unsigned"=true})
     * @ORM\GeneratedValue(strategy="UUID")
     */
    protected $id;

Customer.php
/**
 * @ORM\Entity
 * @ORM\Table(name="customer")
 */
class Customer
{
    /**
     * @ORM\Id
     * @ORM\Column(type="bigint", options={"unsigned"=true})
     * @ORM\GeneratedValue(strategy="AUTO")
     */
    protected $id;

    /**
     * @ORM\OneToOne(targetEntity="User")
     * @ORM\JoinColumn(name="user_id", referencedColumnName="id")
     */
    protected $user;

schema.sql
CREATE TABLE `user` (
  `id` varchar(255) COLLATE utf8_unicode_ci NOT NULL COMMENT '(DC2Type:guid)',
  `name` varchar(60) COLLATE utf8_unicode_ci NOT NULL,
  `username` varchar(40) COLLATE utf8_unicode_ci DEFAULT NULL,
  `username_canonical` varchar(255) COLLATE utf8_unicode_ci DEFAULT NULL,
  `email` varchar(255) COLLATE utf8_unicode_ci NOT NULL,
  `email_canonical` varchar(255) COLLATE utf8_unicode_ci NOT NULL,
  `password` varchar(255) COLLATE utf8_unicode_ci NOT NULL,
  `salt` varchar(255) COLLATE utf8_unicode_ci NOT NULL,
PRIMARY KEY (`id`),
  UNIQUE KEY `UNIQ_8D93D649E7927C74` (`email`),
  UNIQUE KEY `UNIQ_8D93D649F85E0677` (`username`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_unicode_ci;

CREATE TABLE `customer` (
  `id` bigint(20) unsigned NOT NULL AUTO_INCREMENT,
  `user_id` varchar(255) COLLATE utf8_unicode_ci DEFAULT NULL COMMENT '(DC2Type:guid)',
  `gender` varchar(1) COLLATE utf8_unicode_ci DEFAULT NULL,
  `zipcode` varchar(10) COLLATE utf8_unicode_ci DEFAULT NULL,
  `birthday` date DEFAULT NULL,
  PRIMARY KEY (`id`),
  UNIQUE KEY `UNIQ_705B3727A76ED395` (`user_id`),
  CONSTRAINT `FK_705B3727A76ED395` FOREIGN KEY (`user_id`) REFERENCES `user` (`id`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_unicode_ci;





[DDC-2995] Joined Inheritance Mapping and Filters Created: 22/Feb/14  Updated: 14/Oct/14

Status: Open
Project: Doctrine 2 - ORM
Component/s: ORM
Affects Version/s: 2.4.2
Fix Version/s: None

Type: Bug Priority: Major
Reporter: Bojidar Hristov Assignee: Marco Pivetta
Resolution: Unresolved Votes: 0
Labels: orm


 Description   

Is there any reason why Inheritance Mapping IS LEFT JOINED not INNER JOINED?

When there is a filter and it's left joined it happens a record might not have parent table record. For example

Class B extends Class A.
Class A have column is_active | and filter activated with is_active = 1 condition.

Final query: LEFT JOIN parent_table s1_ ON p2_.id = s1_.id AND (s1_.is_active = '1')

Record 1 have is_active = false, so result set looks like this:

table_b | table_a
id | id is_active discriminatorColumn
1 | null null null

and occurs exception: The discriminator column ... is missing for ....



 Comments   
Comment by Marco Pivetta [ 22/Feb/14 ]

JTI MUST be left-joined, since not every subclass of the root of the inheritance causes inserts on all of the inheritance tables.

What is the DQL query that is causing the problem?

Comment by Bojidar Hristov [ 22/Feb/14 ]

Hello Marco, I don't query from parent table, but from child. No sense to be LEFT JOINED.

DQL is something like that:

SELECT class1, class2 FROM Class1 class1 JOIN class1.class2 class2;

(Class2 extends Class3)
Class3 have is_active column.

And following filter:

public function addFilterConstraint(ClassMetadata $targetEntity, $targetTableAlias) {
if (!$targetEntity->reflClass->implementsInterface('ActivateableInterface'))

{ return ""; }

return $targetTableAlias . '.is_active = ' . $this->getParameter('isActive');
}

Filter condition is added in LEFT JOINED parent_table. Full example SQL:
SELECT ....
FROM class1 c1
INNER JOIN class2 c2 ON c1.c2_id = c2.id
LEFT JOIN table3 c3 ON c2.id = c3.id AND (c3.is_active = '1')

Comment by Marco Pivetta [ 26/Feb/14 ]

There are two possible solutions here:

  • verify if the selected entity is a leaf of a JTI, and use INNER JOIN on that if possible (probably not optimal, since the selection may be part of a LEFT JOIN itself)
  • add filtering on the discriminator column

Bojidar Hristov can you provide a failing test case to make it easier to work on this?

Comment by Bojidar Hristov [ 26/Feb/14 ]

Yes. Here is it sample failing test.

http://pastebin.com/xi8uUNX5 - Parent Class
http://pastebin.com/XRZLbyGX - Child Class
http://pastebin.com/HuPFb98u - Assoc Class
http://pastebin.com/hGqRk74e - TestCase

Here it should found 0 records because ot JOIN and is_active = true filter.

Comment by Benjamin Eberlei [ 23/Mar/14 ]

Since filters are only available on the root, i guess inner join is really necessary here.

Comment by Bojidar Hristov [ 14/Oct/14 ]

Do you need something more? The bug is status Feedback?





[DDC-2838] Leaky abstraction when applying Criteria to hydrated/non-hydrated PersistentCollection Created: 03/Dec/13  Updated: 26/Jun/15

Status: Awaiting Feedback
Project: Doctrine 2 - ORM
Component/s: ORM
Affects Version/s: 2.4.1, 2.6.0
Fix Version/s: None

Type: Bug Priority: Major
Reporter: brian ridley Assignee: Marco Pivetta
Resolution: Unresolved Votes: 2
Labels: None

Attachments: File DDC2838Test.php    
Issue Links:
Reference
is referenced by DCOM-275 Collections\Expr\ClosureExpressionVis... Open
is referenced by DCOM-249 Criteria are unable to locate getters... Open

 Description   

When applying a Criteria to a PersistentCollection that has been hydrated the field names must be camel case, if the collection has not yet been hydrated the field names must be underscore separated.

The github repo linked here contains a simplified testcase for the matrix of hydrated/non-hydrated entities and camel case/underscore separated fields.

https://github.com/ptlis/DoctrineTestcase



 Comments   
Comment by Marco Pivetta [ 18/Aug/14 ]

We can't check out an entire project just to test a bug.

Please write an actual functional test case that can be integrated into the Doctrine ORM test suite.

Comment by brian ridley [ 20/Aug/14 ]

Hi,

i'm happy to do so - i'll take a look at this over the weekend.

Comment by Simon Paridon [ 12/Dec/14 ]

brian ridley: Any progress on this? This is currently an issue for us as well and hope to get fixed. I could look into converting it to a test for integration with the test suite if you don't have the time... but it might take a while since I have no experience with the requirements that should be met. (Plus, I am not sure how tightly coupled it is with your project)

Comment by Simon Paridon [ 17/Dec/14 ]

Hi Marco Pivetta, brian ridley,

I attached a functional Test that integrates with the test suite. Please let me know if I should issue a PR, and I'll do that this evening.

Comment by Florian Preusner [ 26/Dec/14 ]

+1

Comment by Simon Paridon [ 26/Jun/15 ]

My idea to solve this would go like this:

  • Add a new class ObjectCollection in Common that implements Collection and Selectable, but requires class metadata.
  • Whenever creating an ArrayCollection with entities / other objects, create an ObjectCollection instead.

The only thing that's causing me a headache is that ideally, there should be code sharing in some form between the matching() implementations of ObjectCollection and PersistentCollection, because both will use class metadata. Maybe this can be achieved somehow using a trait?

If you like the idea, I could look into it further.





[DDC-3832] readOnly should be renamed to immutable in the mapping Created: 16/Jul/15  Updated: 16/Jul/15

Status: Open
Project: Doctrine 2 - ORM
Component/s: Mapping Drivers
Affects Version/s: None
Fix Version/s: 3.0

Type: Improvement Priority: Minor
Reporter: Christophe Coevoet Assignee: Marco Pivetta
Resolution: Unresolved Votes: 0
Labels: None


 Description   

readOnly in the mapping is currently very confusing. Some people are using it to map entities to views, and then complain about SchemaTool-based projects (migrations for instance) because it handles things as table.
But this is not the actual meaning of this mapping flag. Read-only entities can be inserted and deleted. The only forbidden action on them are updates. This makes the naming totally unsuited.

For reference, Hibernate uses "mutable=false" for the feature corresponding to our "readOnly=true".






[DDC-3581] DatabaseDriver does not assert nullable on ToOne associationMappings Created: 21/Feb/15  Updated: 21/Feb/15

Status: Open
Project: Doctrine 2 - ORM
Component/s: Mapping Drivers, Tools
Affects Version/s: 2.4.7
Fix Version/s: None

Type: Improvement Priority: Minor
Reporter: Ryan Korczykowski Assignee: Marco Pivetta
Resolution: Unresolved Votes: 0
Labels: None


 Description   

I am using the DatabaseDriver/AnnotationExporter to tweak class generation and found that JoinColumn annotations did not contain nullable=false even when the local FK column was not null. This is counter-intuitive since nullable on Column annotations is properly handled.

Here is a simple fix that checks the column of the local column in DatabaseDriver::buildToOneAssociationMappings:

for ($i = 0; $i < count($fkColumns); $i++) {
    $associationMapping['joinColumns'][] = array(
        'name'                 => $fkColumns[$i],
        'referencedColumnName' => $fkForeignColumns[$i],
        //fix here
        'nullable' => !$foreignKey->getLocalTable()->getColumn($fkColumns[$i])->getNotNull()
    );
}

Given that the database defines null / not null on FK columns, the JoinColumn nullable attribute should not be left undefined.






[DDC-3527] Provide a method to retrieve exactly one entity out of entity repository Created: 20/Jan/15  Updated: 22/Jan/15

Status: Open
Project: Doctrine 2 - ORM
Component/s: ORM
Affects Version/s: Git Master
Fix Version/s: 3.0

Type: Improvement Priority: Minor
Reporter: Dominik D Assignee: Marco Pivetta
Resolution: Unresolved Votes: 0
Labels: None


 Description   

Hello,

currently we have EntityRepository::findOneBy() method for retrieving a single or none entities out of an EntityRepository. It would be very useful, if there was a method which retrieves exactly one entity or throws NonUniqueResultException in case of multiple records returned or NoResultException in case of zero results returned (just like Query::getSingleResult() currently does).

Very often I find myself creating custom repositories for a couple of ::findOneByOrThrow() methods only. Needless to say it's quite mundane to do. The bodies of such methods follows the same pattern:

1. Query the EntityRepository using EntityRepository::findOneBy().
2. If the result is null, then throw an Exception.
3. Return the result.

It'd be very nice to have it out-of-the-box in the base EntityRepository.

Thanks.



 Comments   
Comment by Marco Pivetta [ 20/Jan/15 ]

We actually want to reduce the repository API, not expand it...

Comment by Dominik D [ 21/Jan/15 ]

Wow, ok. I guess you've got good reasons for it and I shouldn't try to convince you, that having methods for most common use cases out of the box is quite convenient?

If yes, then fine. I can create a common EntityRepository subclass for myself and put any helper methods I want there. This would at least cut down on the code duplication in my codebase.

If what you say is final, then we can close this ticket.

Cheers.

Comment by Marco Pivetta [ 22/Jan/15 ]

I guess you've got good reasons for it and I shouldn't try to convince you, that having methods for most common use cases out of the box is quite convenient?

The main problem is that adding API methods to repositories forces any subclasses to also apply eventual filtering logic to those as well.

Other than that, it's merely a question of interface segregation: the current API is hardly maintainable if we assume that many developers subclassed the EntityRepository class, therefore we should keep it locked until 3.x.

I'll actually mark this issue for 3.0 and defer discussion till then.

Comment by Dominik D [ 22/Jan/15 ]

Ok, thanks.





[DDC-2570] Doctrine CLI Tools - Clear All Cache Created: 24/Jul/13  Updated: 17/Jan/15

Status: Open
Project: Doctrine 2 - ORM
Component/s: ORM, Tools
Affects Version/s: 2.3.4
Fix Version/s: 2.x

Type: Improvement Priority: Minor
Reporter: Frederick Marcoux Assignee: Marco Pivetta
Resolution: Unresolved Votes: 0
Labels: Cli, cache, orm


 Description   

It would be nice to be able to clear all cache one shot instead of clearing them one after one...

Like this:

root$ ./doctrine orm:clear-cache:all

Instead of:

root$ ./doctrine orm:clear-cache:metadata
root$ ./doctrine orm:clear-cache:result
root$ ./doctrine orm:clear-cache:query






[DDC-2227] Add details about developer being responsible of inverse side of an association Created: 09/Jan/13  Updated: 09/Jan/13

Status: Open
Project: Doctrine 2 - ORM
Component/s: Documentation, ORM
Affects Version/s: None
Fix Version/s: None

Type: Documentation Priority: Minor
Reporter: Marco Pivetta Assignee: Marco Pivetta
Resolution: Unresolved Votes: 0
Labels: None


 Description   

As far as I know, docs don't explain that it is up to the developer to keep the object graph consistent instead of relying on Doctrine ORM for everything.

For example, for many to many, examples like following may be used:
https://gist.github.com/3121916






Generated at Sun Aug 30 07:59:20 EDT 2015 using JIRA 6.4.10#64025-sha1:5b8b74079161cd76a20ab66dda52747ee6701bd6.