Let's say I have a Setting entity with some fields like IntValue, dateValue, stringValue and some linked entities, like countries (ManyToMany to entity Country), languages (ManyToMany to Language) etc.
Settings are created by users and assigned to specific objects (not important here, but I wanted to clarify).
Now I suddenly need to have UserDefaultSetting, which will be the same, but with additional user field (ManyToOne to User entity).
I tried to extend existing Setting entity class with one more field added. The problem is, as I looked at the schema update SQL, it created new table for the new entity, but without all the tables needed to ORM connections (mostly ManyToMany). Just one table with "scalar" fields.
So previously I've had setting table with int_value, date_value etc. but also setting_country and setting_language tables, linking ManyToMany relations. After creating child entity, Doctrine created only user_default_setting table with int_value, date_value etc. and additionally user_id column, but I can't see any relation/link tables.
I know I should've been do it with abstract base entity class, but at the time I started, I didn't know that and now part of the project is on production (don't look at me like that, I blame the client) and I don't want to change that "base" class now. Can I inherit everything from non-abstract entity class in a way it will work?
UPDATE: everything explained. See Cerad's comment. Thanks!
Related
I can't believe this hasn't come up for other people, but I'm unable to find a solution.
Let's say I have two entity types, A and B with a one-to-many relationship. A has a collection of Bs.
The form for A has a CollectionType for the Bs, with a custom entry_type for B, allow_add and allow_delete set to true. When the form is created/populated/rendered, the Bs' fields are identified by their index in the collection. When the form is posted back, the fields are mapped back onto the B entities according to the index again.
What if the database in the mean time decided to return the Bs in a different order? Then the values get swapped around on the Bs! I can't have that, as other entities will reference the Bs and now they've changed their meaning!
Even if the database doesn't change the order, the same issue appears when a B is deleted: The fields get shifted through the Bs and a different one deleted! (Ok, I'm not a 100% certain this happens, as there's a gap then in the numbering of the posted fields.) I've found this similar question where it does happen when another one is created (Symfony CollectionType regards deletion+creation as a modification of an item), but that sort of drifted from the issue and there's no usable answer.
How do I make sure the form is updating the entities the user actually edited?
I already tried to render the Bs' IDs as a HiddenType, but then the form rightfully complains that the ID has no setter. It would probably force an ID on the wrong B anyways and Doctrine doesn't like that. I suppose I could add the Bs as unmapped and copy the values over to the correct objects manually, but that would defeat a good chunk of Symfony's form system.
I've used CollectionType before, but not for entities that are referenced elsewhere. I would then delete all of the previous entities and create the collection anew from the posted data. But I can't do that now, can I?
Since doctrine 2.1, it's possible to change how associations are indexed. This will allow you to use the id as the collection key (as the field has to be unique):
#OneToMany(targetEntity="B", mappedBy="A", indexBy="id")
You might also need to enable orphanRemoval so that the data is actually removed instead of the relation just set to null.
I have an existing view in my SQL database "view_account" which represents a account entity. This view is read only and has no primary_key field. Actually it has a primary key "id" but the field is not declared as primary key. I can also not change the table design, because its an automatic export from another application generating this tables (the automatic export is every week).
But thats not the problem cause doctrine don't care about "primary key" flag in the database until you don't update the schema.
But whenever i try to "doctrine:schema:update --force" doctrine wants to create this table. How can i ignore this view (or better the entity) from updating by doctrine? Marking the entity read_only with doctrine annotation is not working. Extending a own update-command will also not work, as i found out, it will not work since doctrine 2.4.7 to extend the doctrine-schema-update command (the update command method is ignored in any way).
Whats the best solution?
Edit:
I also tried to configure two entitymanager. One for the "internal" entities and for the "foreign" entities. Since all entities are linked to each other on some point, it is also not possible to have two separated manager don't knowing from each other entities.
a customer has an existing database. The schema is often changed within the database itself (e.g. he adds a new column).
My task is to develop an admin area with symfony that automatically reacts on table schema changes without modifying the application code. E.g. the customer adds a new column to table "MyEntity", and the application automatically generates a new column in the accordingly list view.
My approach is to dynamically map the table columns to the Entity class so that ALL Attributes and ALL Getters/Setters are generated dynamically from the table schema.
So is it possible to map the table columns in a Doctrine Entity without the use of Annotations or XML Files.
Something like:
class MyEntity{
public function generateMappingFromSchema($sTableName){...}
}
Please don't do that. Doctrine was not designed for such use case.
There is a library though you should check https://github.com/laravel-doctrine/fluent which basically is a mapping driver that allows you to manage your mappings in an Object Oriented approach. And there are other tools:
http://crud-admin-generator.com/
http://crudkit.com/
http://www.grocerycrud.com/
which are maybe better for that, I don't know.
But again, please don't do that. Do not allow the customer to modify the database schema or give them e.g. a phpMyAdmin which was designed for that.
I have entities "Article" and "Product". Now I want to add comments to these 2 entities. Should I create 2 different entities "ArticleComment" and "ProductComment" with the same properties, and build a ManyToOne relation to their respective entity, or create a single "Comment" entity and find a way to build a relation to both "Article" and "Product" entities. Considering solution #2, how could I do that ?
Considering solution #2, how could I do that ?
One way would be to use Single Table Inheritance
Single Table Inheritance is an inheritance mapping strategy where all classes of a hierarchy are mapped to a single database table. In order to distinguish which row represents which type in the hierarchy a so-called discriminator column is used.
This means that you can easily create two separate entities ArticleComment and ProductComment both extending Comment. Then you use the advantages DiscriminatorMap column provides.
Your Comment entity could hold a relation called parent for instance that would refer to either your Article or Product entities. By creating new instance of ArticleComment or ProductComment your discriminator map field would be automatically populated depending on which type you're using.
This would also give you advantages with using DQL to query related comments by their type. Example from documentation:
$query = $em->createQuery('SELECT u FROM Doctrine\Tests\Models\Company\CompanyPerson u WHERE u INSTANCE OF Doctrine\Tests\Models\Company\CompanyEmployee');
And more. You can read that chapter here. Of course this is just a sample and you can use completely different approach.
I've started porting a .NET SQL Server application to LINQ to Entities. I have (among others...) one table called Users, and one called Time. Time is reported on a specific user (UserId), but it is also recorded which user made the report (InsertedByUserId) and possibly who has updated the Time since insert (UpdatedByUserId). This gives me three references to the table Users.
When I generate a .EDMX from this I get three references to the table Users: User, User1 and User2. Without manual edit I have no way of knowing which one refers to the UserId, InsertedByUserId or UpdatedByUserId field.
How do others solve this? Maybe it's not necessary to register ALL references, and stick with InsertedByUserId and UpdatedByUserId as ints?
(The manual edit wouldn't be a problem if the database were never updated, but as we make changes to the database every now and then we occasionally have to regenerate the .EMDX, thus removing all manual changes.)
Thanks in advance!
Jos,
Generally when I make my foreign keys, I name them accordingly. From the Entity designer you can differentiate between the different Navigation Properties (ie User, User1, User2) by looking at the FK association (as long as you named your foreign keys distinctly). For Instance I have a ModifiedById and CreatedById field in each table. Both fields reference my SystemUser table, My foreign keys are named like this: FK_[TableName]_SystemUser_CreatedBy and FK_[TableName]_SystemUser_ModifiedBy.
You should notice that in the Navigation properties you can see the Foreign key. You can also modify the name of the Navigation Property (which is in the Conceptual Side "CSDL portion" of the EDMX), and this change will stay when you update your EDMX from the database.