I am using JMSSerializerBundle to serialize my entities to json and deserialize json into entities, but I think this question applies for any deserialization techniques.
For example, this schema:
class Order
{
private $id;
/**
* #Serializer\Type("ArrayCollection<MyBundle\Entity\Order\Item>")
* #ORM\OneToMany(targetEntity="\MyBundle\Entity\Order\Item", mappedBy="order", cascade={"persist"})
*/
private $items;
}
class Item
{
private $id;
/**
* #ORM\ManyToOne(targetEntity="\MyBundle\Entity\Order", inversedBy="items")
*/
private $order;
/**
* #var integer $amount
* #Serializer\Type("integer")
* #ORM\Column(name="amount", type="integer")
*/
private $amount;
}
Maps to this json: {"id":1,"items":[{"id":1,"amount":100}, {"id":2,"amount":200}]} and the same json is properly deserialized into an object of type MyBundle:Order that has a colletion of two MyBundle:Order/Item objects.
The problem is that when I try to persist this object, new entries are created in the database, rather than updating existing, ignoring the ids. How do I tell entity manager that theses objects should be updated, rather that created?
Update. Generally EntityManager::merge solution (as suggested by DaveM) is fine. But you must only merge existing objects. For example, if you have a json that represents a new Order entity that is connected to existing Order\Item entities
{"id":null,"items":[{"id":1,"amount":100}, {"id":2,"amount":200}]}
In this case you cannot just merge an Order object like this:
$em->merge($order), because order is a new entity and entity manager will attempt to find an Order object with id = null and you will end up with a new Order and empty items array. So the solution is to loop the Order::$items array and merge each item individually. Then a new order will be created and connected with existing items.
You need to use the merge() method on the EntityManager as merging entities refers to the merging of entities into the context of an EntityManager so that they can become managed again. In order to merge the state of an entity into an EntityManager use the EntityManager#merge($entity) method. The state of the passed entity will be merged into a managed copy of this entity and this copy will subsequently be returned.
$detachedEntity = unserialize($serializedEntity);
$entity = $em->merge($detachedEntity);
Also be sure to note when you want to serialize/unserialize entities you have to make all entity properties protected, never private. The reason for this is, if you serialize a class that was a proxy instance before, the private variables won’t be serialized and a PHP Notice is thrown.
More information can be found in the doctrine documentation here:
http://doctrine-orm.readthedocs.org/en/2.0.x/reference/working-with-objects.html#merging-entities
I know this question is three years old, but it mislead me to think the only answer was using the merge operation. I'd like to add my two cents:
The JMSSerializerBundle includes an object constructor for Doctrine entities. When you enable this constructor, the deserialized entities are managed entities that can be persisted(with $em->persist($entity)).
Please check this comment to understand other benefits from this.
And here is how you can enable it.
Related
I am getting an error:
Property or Indexer cannot be assigned to "--" it is read only
when trying to update two columns with the same name in two tables in a join query. How do I get this to work? Thanks!
The anonymous object created in your projection ("select new" part) is read-only and its properties are not tracked by data context by any means.
Instead, you can try this:
//...
select new
{
p1 = p,
p2 = t
}
foreach (var row in updates)
{
row.p1.Processed = true;
row.p2.Processed = true;
}
In order to improve performance you may also want to take a look at batch update capabilities of Entity Framework Extensions (if you are using Entity Framework): https://entityframework-extensions.net/overview
Yes, that's due to anonymous type properties are read only, from documentation:
Anonymous types provide a convenient way to encapsulate a set of
read-only properties into a single object without having to explicitly
define a type first.
I suggest you to create a custom class with the two entities you need (a DTO):
public class PassengerDTO
{
public Passenger Passenger {get;set}
public PassengerItinerary PassengerItinerary {get;set}
}
And use it in your projection, You need the entity instances and not just the properties you want to modify because, when you modify the Processed property in the foreach the proxy class that represent your entity is going to change the status of you entity to Updated.
This is classic method for the entity relationship;
/**
* #ORM\OneToMany(targetEntity="Product", mappedBy="category")
*/
protected $products;
and other entity;
/**
* #ORM\ManyToOne(targetEntity="Category", inversedBy="products")
* #ORM\JoinColumn(name="category_id", referencedColumnName="id")
*/
protected $category;
But, when I use this method, I can't use limit function. All results are taken. But, I should use sql limit function. So, I use a repository class. I wrote all sql code again. I used join function again. Is the relationsip annotation necessary in this situation? Do I still need to add this annotation?
Unfortunately, there is no way of doing what are you asking. You are gonna have to write a custom DQL, preferably inside a custom repository in order to fetch exactly what you want. Of course you still need the annotation itself, as is required by Doctrine to do many things under the hood and, in general, to keep a good relationship model between your entities.
The most close annotation you have to control the way in which relationships are joined when using automatic method is the OrderBy one:
/*
* #OneToMany(...)
* #OrderBy({"name" = "ASC"})
*/
protected $relation;
When I load an entity which contains lazy-loading fields and I want to send this entity through an object message with ActiveMQ, will I receive :
The full entity (with lazy-loading fields loaded)
OR
The entity as sent (without lazy-loading fields loaded) ?
In any case do I need to put the Serializable marker in my entity ?
In case of answer 1 what do I need to do to get the entity as described in answer 2 ?
Lets assume you have an entity with some lazy loading fields like this:
#Entity
public class Foo
{
#Id
private Long id;
#OneToMany
private List<Bar> bars;
}
When you receive this entity from your database only the id field will be loaded as the bars field is lazy loaded (OneToMany is lazy by default). Now when you pass this entity through JMS for example the bars field is not initialized because the getBars() method isn't called. When the remote end calls the getter it will get a LazyInitializationException as the entity is detached and the collection isn't initialized.
If you however want the remote end to be able to call the getBars() method you need to initialize the collection in some way. You can make the fetching of the collection eager using #OneToMany(fetch = FetchType.EAGER). An alternative is to use a separate query for eager loading the entity (which is my personal preference).
Example:
select f from Foo f join fetch l.bars
If you really want to go wild and want the collection to be initialized lazily on the remote end you can take a look at Hibernate remote lazy loading using RMI. I'm not going to explain this because it seems out of the scope of the question. ;-)
Currently I am working partly with cfwheels and its Active Record ORM (which is great), and partly raw cfml with its Hibernate ORM (which is also great).
Both work well for applicable situations, but the thing I do miss most when using CF ORM is the model.update() method that is available in cfwheels, where you can just pass a form struct to the method, and it will map up the struct elements with the model properties and update the records.. really good for updating and maintaining large tables. In CF ORM, it seems the only way to to update a record is to set each column individually, then do a save. Is this the case?
Does cf9 ORM have an Active Record type update() (or equivalent) method which can just receive a struct with values to update and update the object without having to specify each one?
For example, instead of current:
member = entityLoadByPK('member',arguments.id);
member.setName(arguments.name);
member.setEmail(arguments.email);
is there a way to do something like this in CF ORM?
member = entityLoadByPK('member',arguments.id);
member.update(arguments);
Many thanks in advance
In my apps I usually create two helper functions for models which handle the task:
/*
* Get properties as key-value structure
* #limit Limit output to listed properties
*/
public struct function getMemento(string limit = "") {
local.props = {};
for (local.key in variables) {
if (isSimpleValue(variables[local.key]) AND (arguments.limit EQ "" OR ListFind(arguments.limit, local.key))) {
local.props[local.key] = variables[local.key];
}
}
return local.props;
}
/*
* Populate the model with given properties collection
* #props Properties collection
*/
public void function setMemento(required struct props) {
for (local.key in arguments.props) {
variables[local.key] = arguments.props[local.key];
}
}
For better security of setMemento it is possible to check existence of local.key in variables scope, but this will skip nullable properties.
So you can make myObject.setMemento(dataAsStruct); and then save it.
There's not a method exactly like the one you want, but EntityNew() does take an optional struct as a second argument, which will set the object's properties, although depending on how your code currently works, it may be clunky to use this method and I don;t know whether it'll have any bearing on whether a create/update is executed when you flush the ORM session.
If your ORM entities inherit form a master CFC, then you could add a method there. Alternatively, you could write one as a function and mix it into your objects.
I'm sure you're aware, but that update() feature can be a source of security problems (known as the mass assignment problem) if used with unsanitized user input (such as the raw FORM scope).
It seems that serializing Entity Framework objects into JSON is not possible using either WCF's native DataContractJsonSerializer or ASP.NET's native JavaScript serializer. This is due to the reference counting issues both serializers reject. I have also tried Json.NET, which also fails specifically on a Reference Counting issue.
Edit: Json.NET can now serialize and deserialize Entity Framework entities.
My objects are Entity Framework objects, which are overloaded to perform additional business functionality (eg. authentication, etc.) and I do not want to decorate these classes with platform-specific attributes, etc. as I want to present a platform-agnostic API.
I've actually blogged about the individual steps I went though at https://blog.programx.co.uk/2009/03/18/wcf-json-serialization-woes-and-a-solution/
Have I missed something obvious?
The way I do this is by projecting the data I want to serialize into an anonymous type and serializing that. This ensures that only the information I actually want in the JSON is serialized, and I don't inadvertently serialize something further down the object graph. It looks like this:
var records = from entity in context.Entities
select new
{
Prop1 = entity.Prop1,
Prop2 = entity.Prop2,
ChildProp = entity.Child.Prop
}
return Json(records);
I find anonymous types just about ideal for this. The JSON, obviously, doesn't care what type was used to produce it. And anonymous types give you complete flexibility as to what properties and structure you put into the JSON.
Microsoft made an error in the way they made EF objects into data contracts. They included the base classes, and the back links.
Your best bet will be to create equivalent Data Transfer Object classes for each of the entities you want to return. These would include only the data, not the behavior, and not the EF-specific parts of an entity. You would also create methods to translate to and from your DTO classes.
Your services would then return the Data Transfer Objects.
Based off of #Craig Stuntz answer and similar to a DTO, for my solution I have created a partial class of the model (in a separate file) and a return object method with how I want it using only the properties that will be needed.
namespace TestApplication.Models
{
public partial class Employee
{
public object ToObject()
{
return new
{
EmployeeID = EmployeeID,
Name = Name,
Username = Username,
Office = Office,
PhoneNumber = PhoneNumber,
EmailAddress = EmailAddress,
Title = Title,
Department = Department,
Manager = Manager
};
}
}
}
And then I call it simply in my return:
var employee = dbCtx.Employees.Where(x => x.Name == usersName).Single();
return employee.ToObject();
I think the accepted answer is more quick and easy, I just use my method to keep all of my returns consistent and DRY.
My solution was to simply remove the parent reference on my child entities.
So in my model, I selected the relationship and changed the Parent reference to be Internal rather than Public.
May not be an ideal solution for all, but worked for me.
One more solution if you want to have better code consistency is to use JavaScriptConverter which will handle circular reference dependencies and will not serialize such references.
I've blogged about here:
http://hellowebapps.com/2010-09-26/producing-json-from-entity-framework-4-0-generated-classes/
FYI I found an alternative solution
You can set the parent relationship as private so then the properties are not exposed during the translation removing the infinite property loop
I battled with this problem for days,
Solution. Inside your edmx window.
- right click and add code generation item
- Select Code tab
- select EF 4x.POCOC Entity Generator
If you don't see it, then you will have to install it with nuget, search EF.
The Entity generator will generate all you complex type and entity object into simple classes to serialize into json.
I solved it by getting only object types from System namespace, and then convert them to Dictionary and then add them to list. Works good for me :)
It looks complicated, but this was the only generic solution that worked for me...
I'm using this logic for a helper I'm making, so it's for a special use where I need to be able to intercept every object type in entity object, maybe someone could adapt it to his use.
List<Dictionary<string, string>> outputData = new List<Dictionary<string, string>>();
// convert all items to objects
var data = Data.ToArray().Cast<object>().ToArray();
// get info about objects; and get only those we need
// this will remove circular references and other stuff we don't need
PropertyInfo[] objInfos = data[0].GetType().GetProperties();
foreach (PropertyInfo info in objInfos) {
switch (info.PropertyType.Namespace)
{
// all types that are in "System" namespace should be OK
case "System":
propeties.Add(info.Name);
break;
}
}
Dictionary<string, string> rowsData = null;
foreach (object obj in data) {
rowsData = new Dictionary<string, string>();
Type objType = obj.GetType();
foreach (string propertyName in propeties)
{
//if You don't need to intercept every object type You could just call .ToString(), and remove other code
PropertyInfo info = objType.GetProperty(propertyName);
switch(info.PropertyType.FullName)
{
case "System.String":
var colData = info.GetValue(obj, null);
rowsData.Add(propertyName, colData != null ? colData.ToString() : String.Empty);
break;
//here You can add more variable types if you need so (like int and so on...)
}
}
outputData .Add(rowsData); // add a new row
}
"outputData " is safe for JSON encode...
Hope someone will find this solution helpful. It was fun writing it :)