we have a very dynamic metamodel which changes frequently. We don't use EF but an own Database-Technology.
As we didn't find any good documentation on how the server-side metamodel has to be created "by hand", we decided to create the client-side metamodel and import this one into breeze when loading the application.
Now we have the problem, that eventhough we tell breeze to get all Person-Objects and take 3, the server expands all associations and returns more than 3 results. Breeze then cuts the result on the client to 3 and fills the properties.
Now I'm wondering, if we just didn't understand some of the concepts. All we want to do is to tell breeze dynamically how our metamodel looks like without using EF or NHybernate. Is there any documentation on that?
The only documentation on a Schema I found was this one:
http://www.breezejs.com/documentation/metadata-schema
But it only explains the client-Side-Metamodel but not the Schema that our Server-Side would have to generate. As far as I see it, the Server-Side-Metamodel gets translated into the client-Side-Metamodel anyways.
Would be nice if someone could clarify or provide a link with all the "basic information" about the topic.
The client metadata is for the Breeze client. The server shouldn't need it, because it already knows what the domain model looks like. The content of server-side metadata typically comes from the database and/or the mapping layer, when using an ORM.
Breeze's server-side WebApi filters attempt to apply the OData query parameters to the IQueryable that comes from the data provider (EF, NH, or whatever). If the data provider's LINQ implementation is incomplete, that could cause the problem you mentioned, in which take didn't work.
Unfortunately, the metadata-schema document is out-of-date. The structure of the Breeze metadata JSON format has changed since then. You are better off looking at an example of current metadata, and following the Metadata by Hand guidelines.
Related
I just want to know how Entity Framework internally works to reveal properties and their types in runtime, particularly in case of Code-First approach, where there won't be system generated code. Can some body give some heads up? I don't think System.Reflection was being used implicitly?
Code first was first presented to developers as part of the EF Feature
CTP1 in June 2009 with the name “code only.” The basic premise behind
this variation of using the EF was that developers simply want to
define their domain classes and not bother with a physical model.
However, the EF runtime depends on that model’s XML to coerce queries
against the model into database queries and then the query results
from the database back into objects that are described by the model.
Without that metadata, the EF can’t do its job. But the metadata does
not need to be in a physical file. The EF reads those XML files once
during the application process, creates strongly typed metadata
objects based on that XML, and then does all of that interaction with
the in-memory XML.
Code first creates in-memory metadata objects, too. But instead of
creating it by reading XML files, it infers the metadata from the
domain classes (see Figure 1). It uses convention to do this and then
provides a means by which you can add additional configurations to
further refine the model.
ModelBuilder will now take this additional information into account as
it’s creating the in-memory model and working out the database schema.
By Julie Lerman
In our entity framework model that identify the customer, a simple Customer_GUID. We are using breeze with asp.net mvc and doing IQueryable.
Is there a way to globally not return those columns in the JSON? This would reduce a good bit of data coming across the wire. We don't want to remove it from mapping in our EF model because we still use it when we save.
You might want to look at the Json.NET documentation in particular the [JsonIgnore] attribute. Look at "Conditional Property Serialization" for more sophisticated scenarios.
Do be careful about insert and update data coming from the client. You'll have to do something if your client uploads a new entity for insertion and it lacks the properties you require on the server side.
To be clear, your configuration of Json.NET has no affect on your server-side EF model ... exactly what you wanted.
That also means that metadata generated from your EF model will describe properties the client can't see. You'll want to compensate for that I imagine. Such compensation is beyond the scope of this question; look to the Breeze documentation on metadata ... particularly "Metadata by hand" and "EF as a design tool".
This is actually reposting a question that already exists but I don't think it was properly understood and for us is really important to know if is possible or if it will be:
https://stackoverflow.com/questions/16079703/how-would-one-go-with-saving-a-complex-object-graph-as-xml-in-sql-database-whil
So, what we would like to know is how do we get to transfer from Breeze to server NON MAPPED TO DB entities/properties. For example, let's consider XML (I would't want to generate xml in JS, but I do have XML db columns that need to be populated from complex forms - so we will collect the data in Breeze/KO, will transfer it to server and on server will process and generate the XML, from the NON MAPPED Entities/Properties).
P.S.
I see there is already a NODB approach (http://www.breezejs.com/samples/nodb), so would be really nice if we would be able to make the 2 approaches work together (EF + NODB)
As of Breeze v 1.3.6, there is now an EntityInfo.UnmappedValuesMap property available during the save that exposes all of the unmapped properties on any entity being saved.
Providing I'm understanding your question correctly, any properties declared as 'unmapped' on a breeze entity do get transferred to the server on a save for exactly this purpose. You can intercept and work with this data within the server side BeforeSaveEntity and BeforeSaveEntities methods.
There is more info here regarding "unmapped" properties:
http://www.breezejs.com/documentation/extending-entities
Let's say I have an entity called Product and this entity is loaded every time user hits the product information page. Usually I'd save the object in Zend_Cache (memcache) for an hour to avoid hitting the db for each request but as far as I understand that's not possible with Doctrine2 entities because of the Proxy objects.
So my question is, how can I avoid loading the same entity from the database for each request?
[EDIT]
I tried using Doctrine Cache like this
$categoryService = App_Service_Container::getService('\App\Service\Category');
$cache = $categoryService->getEm()->getConfiguration()->getResultCacheImpl();
$apple = $cache->fetch('apple');
But I get the following error
Warning: require(App/Entity/Proxy/_CG_/App/Entity/Category.php)
[function.require]: failed to open stream: No such file or directory
in /opt/vhosts/app/price/library/Doctrine/Common/ClassLoader.php on
line 163
This is same for Zend Cache as well as you can't serialize the entity because of the Proxy class
You've got several options:
Use Doctrine's built-in result caching
Try just sticking entity in memcache via Zend_Cache. When you pull it out, you may need to merge() the Product back into the EM so proxies can be dereferenced. If you fetch-join any associations you need to display the product info, and you're only doing reads, this shoudl work fine.
Don't cache the entity at all. Cache whatever output you generate instead.
EDIT: If you don't care about the hydration overhead, you're using mysql, and your Products and associated tables don't change very often, you might prefer to just rely on the mySQL query cache. It's a fairly blunt object, but useful enough to mention.
You might want to try implementing __sleep or __wakeup methods for your entity class, as Doctrine 2 has special requirements and limitations concerning serialization/deserialization of entities (which is what happens when storing them in Zend_Cache).
There is this guidance.
General information about limitations including serialization.
I find this extremely strange since i just messed around with this myself and didn't have any issues with the proxy object being stored in the database. So im guessing your configuration is not setup 100% ?
If you find the issue with your configuration then be very aware of what timdev said you MUST merge the object back into the EntityManager else you will have weird bugs down the line.
A fourth solution available for you is also to retrieve the data as an array instead of an object, but then of course you lose all the functionality connected to your module which might not be exactly want you wanted.
It seems to me more like a configuration error. Either Proxies have not been generated or there is something wrong with the proxy directory and namespace.
Depending on your configuration, proxies can be either generated automatically or manually. Does your proxies have been indeed generated under App/Entity/Proxy ? Is this indeed the right directory?
FYI proxies can be manually generated by executing doctrine orm:generate-proxies <dest-dir>
Seconding what timdev says: Doctrine has built-in caching, you want to use it.
I also wonder from your question if you are experiencing any performance issues or if you are a victim of overly eager optimisation.
I have now decided to try out ASP.NET MVC 3.
My host provider, however, only supports MySQL and therefore I have to figure out how to use MVC 3 with MySQL.
I have also decided that I don't wanna do any SQL code if I can avoid it, and I would also like O/RM without too much effort. I understand that the Entity Framework will actually help me accomplish this to a large extent.
I have been trying to get into the various ways of using the EF, with the database first, model first and code first approaches supplied by the framework.
So far, I have not had much luck, and I find that the examples available all use very different approaches that confuses me a lot.
I might begin by asking for guidance on getting a few concepts right.
First of all, the Model (in MVC) is actually more like a ViewModel, that represents something (Users, Posts, etc.) in terms of Properties is more or less simple classes. I.e. the model is where the data from the database gets mapped to an object (the O/RM). Am I right?
A repository is a wrapper that encapsulates a specific way of retrieving data for the models. For instance, a DatabaseRepository or a FakeTestRepository.
Should I have a single repository in my MVC project, or a repository per database table, such that I have a UsersRepository and PostsRepository?
Should the repository be a model for itself, not a model at all, or tied to individual models (so that UsersRepository is part of the UsersModel)?
I have tried to use the EF's model first approach, and for a simple test I just have created an empty model and added the entities "Author" and "Guide" that are related by a one-to-many relation.
When I then, in Visual Studio 2010, "Generate database from model", I get the corresponding sql code. I want this database to be created in MySQL. How can I accomplish that?
Are there some code examples for MVC 3 with MySQL and O/RM where the creation of a small site is demonstrated?
Thanks.
Concerning EF Model First approach: take a look at this Tips & Tricks article. We have described this common situation in it (it is Oracle-specific, but dotConnect for MySQL contains the "Devart SSDLToMySQL.tt" template).
As for the rest of the questions - there is no definite answer. Choose the approach that suits you better.
In my point of view, you should try the code first. And as you said that your host only provides MySQL you can also use MySQL database as a database I personally use MySQL. Concepts are the same but logic is different you have to code it a different way. But from my point of view, you can use MySQL as a database service.