[DDC-3764] MappingException thrown on table not in filter Created: 11/Jun/15  Updated: 12/Jun/15

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

Type: Bug Priority: Minor
Reporter: Jordan Gigov Assignee: Benjamin Eberlei
Resolution: Unresolved Votes: 0
Labels: import, mapping, postgresql


 Description   

I'm using the command line to import the database schema from en existing project in another language, however this happens every time

jordan@jordan:~/workspace/testimport$ app/console -v doctrine:mapping:import OldBundle --filter=Users

  [Doctrine\ORM\Mapping\MappingException]
  It is not possible to map entity 'QrtzTriggers' with a composite primary key as part of the primary key of another entity 'QrtzCronTriggers#schedName'.
  

As maybe you can see I've specified a filter to only give me one table, but it throws an exception on a completely unrelated one. There are absolutely no database relations between the Qrtz tables and everything else. The only workaround is to drop the tables, just so I can do the import, but that means I'll have to restore the database later. Those tables are for a distinct module within our system, which can't be ported, even if we decide to migrate.

Whatever analyser you're calling in there should really respect the filters. Not only is it doing extra work, but it's also causing this crash.



 Comments   
Comment by Marco Pivetta [ 11/Jun/15 ]

Loading invalid mappings, regardless of how they are filtered afterwards, is still invalid.

Comment by Jordan Gigov [ 12/Jun/15 ]

Invalid? Tell that to the people at Postgre and Quartz. I'm sure they'll agree you can't possibly have an entity with a composite primary key be referenced by others.

Anyway I'm not asking for composite key references to be implemented (I could try doing it once I get familiar with the system), I'm asking that the filter be respected.

Comment by Marco Pivetta [ 12/Jun/15 ]

Invalid? Tell that to the people at Postgre and Quartz.

Well, the ORM, like any other tool, has its limits.

The filtering happens post-load, so your suggestion would require a massive rewrite of the entire functionality.

You could also simply dump the relevant schema and drop the tables from the dump, as you suggested above.





[DDC-3333] doctrine:schema:update --complete does not detect old index Created: 02/Oct/14  Updated: 02/Oct/14

Status: Open
Project: Doctrine 2 - ORM
Component/s: Tools
Affects Version/s: 2.4.4
Fix Version/s: None

Type: Bug Priority: Minor
Reporter: Grégoire Paris Assignee: Benjamin Eberlei
Resolution: Unresolved Votes: 0
Labels: postgresql, schematool
Environment:

Ubuntu 14.04, PostgreSQL 9.3



 Description   

When changing the name of an index and using the symfony proxy command doctrine:schema:update --complete --dump-sql, the output only shows a CREATE statement. I would expect to also see a DROP statement. Here is what my index definition looks like :

    indexes:
        admin_entity_created_at_index:
            columns: [ createdAt ]





[DDC-3141] Change the type of a column Created: 28/May/14  Updated: 28/May/14  Resolved: 28/May/14

Status: Resolved
Project: Doctrine 2 - ORM
Component/s: DQL, ORM, Tools
Affects Version/s: 2.x
Fix Version/s: None

Type: Improvement Priority: Minor
Reporter: Mathias STRASSER Assignee: Marco Pivetta
Resolution: Won't Fix Votes: 0
Labels: postgresql
Environment:

All O.S.



 Description   

Is that possible to change the request when we would like to change the type of a column in PostgreSQL via Symfony 2 ?

app/console doctrine:schema:update --force

[Doctrine\DBAL\DBALException]
  An exception occurred while executing 'ALTER TABLE foo ALTER valid TYPE BOOLEAN':  

  SQLSTATE[42804]: Datatype mismatch: 7 ERROR:  column "valid" cannot be cast to type boolean

For resolve this problem I must execute manually this command :

ALTER TABLE foo ALTER valid TYPE BOOLEAN USING valid::BOOLEAN;

I've asked this question to DoctrineBundle, but they redirected to you (https://github.com/doctrine/DoctrineBundle/issues/292)



 Comments   
Comment by Marco Pivetta [ 28/May/14 ]

Mathias STRASSER does this happen also when data is not in the table?

To be honest, this exception seems valid to me, as it actually prevents you from applying destructive DDL on your schema.

Comment by Steve Müller [ 28/May/14 ]

I tend to agree with Marco Pivetta. I know of similar issues with other vendors, too and IMO this exception is perfectly valid as you would otherwise risk to loose data integrity. How would you expect a database to do data conversion between non-compatible types? Maybe you can force type conversion for some types but I don't think it is the task of DBAL to handle this. You should handle this manually IMO.

Comment by Mathias STRASSER [ 28/May/14 ]

I understand this prevent but I think if a user asked to apply a schema update, it should be applied or he should get prompted for a confirmation.
It think it's a pity we are forced to manually perform this query if we are sure.

What do you think ?

Comment by Steve Müller [ 28/May/14 ]

Mathias STRASSER I understand your concern. But this is an exception coming from the database server. How would you expect Doctrine to do the data conversion? Maybe in your case when converting to a BOOLEAN column, this could somehow work (even at database level). But how would you for example expect Doctrine to convert a STRING column to INTEGER for example. You just can't do that. Also you have to keep in mind that DBAL is an abstraction layer and thus requires a cross-vendor behaviour that is the same for all. It might be that PostgreSQL can force the conversion of any column type to BOOLEAN but other vendors can't do this natively. You could adopt the scenario to PHP and type casting. It's a similar scenario.

Comment by Marco Pivetta [ 28/May/14 ]

Won't be fixed.

As I've previously noted, this is an engine error that simply avoids a lossy conversion.

The conversion should be forced manually by the person running the schema tool (by manually running the dump command and applying the DDL statements)

Comment by Mathias STRASSER [ 28/May/14 ]

I quite understand.

Thanks for your replies.





[DDC-2424] Removing an inherited entity via a delete cascade constraint does not remove the parent row Created: 02/May/13  Updated: 15/Oct/14

Status: Awaiting Feedback
Project: Doctrine 2 - ORM
Component/s: None
Affects Version/s: 2.3.3, 2.4.6
Fix Version/s: None

Type: Bug Priority: Major
Reporter: Bruno Jacquet Assignee: Benjamin Eberlei
Resolution: Unresolved Votes: 3
Labels: inheritance, postgresql
Environment:

Mysql 5.1.66 / Symfony 2.2.1



 Description   

For a parent class:

/**
 * @ORM\Entity
 * @ORM\Table(name="Base")
 * @ORM\InheritanceType("JOINED")
 * @ORM\DiscriminatorColumn(name="discr", type="string")
 * @ORM\DiscriminatorMap({"child1" = "Child1", "child2" = "Child2"})
 */

and simple Child1 & Child2 entities.

With another entity (let's call it ExternalEntity) having a bidirectional OneToOne relation owned by Child1:

class Child1 extends Base
{
  /**
   * @ORM\OneToOne(targetEntity="ExternalEntity", inversedBy="xxx")
   * @ORM\JoinColumn(onDelete="CASCADE", nullable=false)
   */
   private theForeignKey;
}

Enough for the context.
The symptoms:

$em->remove(instanceOfExternalEntity);

removes the ExternalEntity row and the Child1 row. But a dangling row in the Base table is still there for the now inexistent Child1 instance.

Though, a manual delete of either the associated Child1 OR Base row and then the ExternalEntity works.

The problem with the cascading deletion of the parent seems to be only present when deleting through a MYSQL cascading delete from another row which has a foreign key on a child. (Not tested with a foreign key on the parent though)



 Comments   
Comment by Benjamin Eberlei [ 04/May/13 ]

Can you show the CREATE TABLE and FOREIGN KEY statements of all the tables involved? It seems the cascade of the foreign keys is not propagated between multiple tables?

Comment by Bruno Jacquet [ 06/May/13 ]

CREATE TABLE Base (id INT AUTO_INCREMENT NOT NULL, PRIMARY KEY(id)) DEFAULT CHARACTER SET utf8 COLLATE utf8_unicode_ci ENGINE = InnoDB;
CREATE TABLE Child1 (id INT NOT NULL, foreignKey INT NOT NULL, UNIQUE INDEX UNIQ_179B6E88E992F5A (foreignKey), PRIMARY KEY(id)) DEFAULT CHARACTER SET utf8 COLLATE utf8_unicode_ci ENGINE = InnoDB;

ALTER TABLE Child1 ADD CONSTRAINT FK_179B6E88E992F5A FOREIGN KEY (foreignKey) REFERENCES ExternalEntity (id) ON DELETE CASCADE;
ALTER TABLE Child1 ADD CONSTRAINT FK_179B6E8BF396750 FOREIGN KEY (id) REFERENCES Base (id) ON DELETE CASCADE;

Comment by Bruno Jacquet [ 06/May/13 ]

The problem is that, the SQL model never explicitely tells the DB to delete the corresponding Base when Child1 gets removed. It looks like it is handled by the doctrine entity manager layer and not the actual DB engine (Base has no on delete cascade nor foreign key to its children).
So only doctrine can add the logic here because it knows the entity schema. But in this case, when it is deleted from another table, it looks like the special treatment is not triggered.

Comment by Bruno Jacquet [ 06/May/13 ]

Maybe using

cascade={"remove"}

, instead of

onDelete="CASCADE"

to force the cascading process to be handled by doctrine would workaround the bug... But I prefer to have my DB do the logic work as much as possible.

Comment by J [ 25/Apr/14 ]

I've got a similar problem but I have InheritanceType("SINGLE_TABLE") instead of JOINED.
Any updates on when this is getting fixed?





[DDC-2238] doctrine:schema:update partially broken Created: 11/Jan/13  Updated: 01/May/13  Resolved: 01/May/13

Status: Resolved
Project: Doctrine 2 - ORM
Component/s: Tools
Affects Version/s: 2.3
Fix Version/s: None

Type: Bug Priority: Major
Reporter: Tom Vogt Assignee: Benjamin Eberlei
Resolution: Duplicate Votes: 0
Labels: postgresql, schematool
Environment:

OS X 10.7.5
PHP 5.4.4
Postgres SQL 9.1.6

also confirmed on:
Linux (Debian testing, Kernel 2.6.32)
PHP 5.4.4
PostgreSQL 9.1.9
Doctrine 2.3.2


Attachments: Text File schemadiff.txt    
Issue Links:
Duplicate
duplicates DBAL-504 DBAL Enum fields migration issue / Po... Resolved

 Description   

the app/console doctrine:schema:update command generates a seemingly random number of statements like these:

ALTER TABLE geodata ALTER humidity SET ;
ALTER TABLE geodata ALTER lake SET ;
ALTER TABLE message ALTER translate SET ;

which are obvious invalid SQL commands. The mappings are fine, validate and the application works just fine. Here's an example from the mapping files including two of the above statements:

<field name="coast" type="boolean"/>
<field name="lake" type="boolean"/>
<field name="river" type="boolean"/>
<field name="humidity" type="float"/>
<field name="biome" type="string"/>

I am using doctrine2-spatial as an extension for GIS information. This problem shows up both in entities using spatial data and entities not using spatial data.

I'll gladly help debug this, as right now I can't update my dev database with --force, I need to use --dump-sql and filter out the invalid lines.



 Comments   
Comment by Benjamin Eberlei [ 12/Jan/13 ]

Can you dump the SchemaDiff/ColumnDiff instances that are returned from lib/Doctrine/DBAL/Schema/Comparator.php?

Comment by Tom Vogt [ 12/Jan/13 ]

requested dump of SchemaDiff

Comment by Tom Vogt [ 12/Jan/13 ]

added in the file schemadiff.txt - I added this in Schematool.php - getUpdateSchemaSql():

$comparator = new Comparator();
$schemaDiff = $comparator->compare($fromSchema, $toSchema);
\Doctrine\Common\Util\Debug::dump($schemaDiff, 6);

if you need other output, just tell me what entity to dump and where. I'll be happy to help.

The file also contains the buggy update statements it creates towards the end. There's a few non-crucial bugs included, where it alters the geospatial columns to themselves (i.e. river.course already is a geometry/linestring). I don't worry about those because they don't break anything.

Comment by Tom Vogt [ 15/Apr/13 ]

Is there any update on this? I'm still having this issue, with many different entities, always the same problem, for example:

ALTER TABLE event ALTER priority SET ;

Which is an integer field on an entity that doesn't have any GIS elements, so I'm not even sure if it's caused by that anymore.

Comment by Benjamin Eberlei [ 01/May/13 ]

Tom Vogt I added a fix for PostgreSQL today, can you verify again if this works? Its included in the 2.3 branch.

Comment by Benjamin Eberlei [ 01/May/13 ]

Duplicate of http://www.doctrine-project.org/jira/browse/DBAL-504





Generated at Sat Aug 29 03:28:32 EDT 2015 using JIRA 6.4.10#64025-sha1:5b8b74079161cd76a20ab66dda52747ee6701bd6.