Monday, October 01, 2007

One of the biggest challenges with Entity Framework for updates is that if you want to move objects across tiers, you lose the object's state. All of the state and change tracking information is maintained by the object context. While entity objects are serializable (xml and binary) and can be transported across tiers, object context is not.

 

Therefore, if you have objects being managed  by an ObjectContext in one tier, then move them to another tier and attach them to a new ObjectContext, you will have the current data, but nothing else and the object will be seen as "unchanged".

 

If you want to use the ObjectContext's SaveChanges method to do your db updates (I am using the term "update" generically to refer to updates, inserts and deletes, aka "the CUD in CRUD"), nothing will get sent to the server since it's all seen as unchanged.

 

There are a few ways we have seen to get around this., but they have all boiled down to this:

 

Solution #1 Ignore concurrency and perform an inefficient update

 

Buried in replies by team members in the ADO.NET Orcas forums there is a  conceptual solution that leverages the EntityState object (combined with EF's MetaDataWorkSpace or reflection) to mark the object as modified and then iterate through ALL of the properties in your object and change their state to "modified". Now SaveChanges will push all of the current data up to the server.

 

What I like about this solution:
It's reusable code that will work with any entity which you throw at it.

 

What I don't like about this solution:

1) What is hugely unsatisfying to me is that this is a pure "last data in wins" scenario. While this is probably all that many solutions require, it means that there is no interest in handling any possible conflicts for data that may have changed on the server in between the user's fetch and subsequent save.

2) Every property is flagged for update. Every property value is sent to SQL Server and every property value is updated in the database. Icch.

 

Solution #2 Write object update code for every single entity class

Another workaround I came up with was to extend the entity class so that it knows how to update itself. This assumes we have kept the original data hanging around somewhere. For each class, I created a method that takes in the current entity and the original entity, then I have some code that explicitly updates fields that I care about updating. Here's part of that, where I'm dealing with updates, but not deletes and inserts.

 

 

This works - the original order now has the current values updated to the new values and the state is impacted correctly for the properties and the entity itself. Then I can use SaveChanges or other functionality that I've exposed in my model such as stored procedures to get the data to the server. The key here is that I can leverage the framework for my DML.

 

But this solution stinks because I have to write the same method over and over again for every single class.

 

Solution that I want and finally figured out: Update only relevant properties and have the option to deal with concurrency on a granular level

Mike  Taulty actually pulled this off in this blog post, however I explicitly chose to hold off on reading Mike's post. :-) I didn't want to be influenced by his methods. What's interesting is that in the end, we came up with very similar solutions, even though I went down many many paths before I found success! (I'm sure it took me about 10 times as long as it took Mike though.) Some evidence of my doggedness is in Dave Burke's post about what an un-fun girl I was on Saturday night at code camp.

 

Updating for efficiency and concurrency

 

So here's what I came up with as part of a larger solution of using entity framework in an ASP.NET website, surviving post backs, keeping the original values around and being able to use SaveChanges and also for using entities across tiers (eg web services).

 

I use reflection and the Entity Framework MetaDataWorkspace to accomplish this.

 

I'll write another post about how I'm keeping the original values and surviving the postbacks on the ASP.NET site.

 

#1 important rule that has been discovered by everyone who is trying to solve this problem in Entity Framework and in LINQ to SQL, use a datetime field in your entity. This can be a field that is already in your database table or a special property exposed only by the conceptual model (or an extension to the entity's class). If it's in the database, you're original data is fine, but you need to be sure to update that field any time you make mods to the properties of the entities. If you create a special property in the EDM, then you'll also need to set the date when you first retrieve data from the data store.

 

Below, I am starting with a collection of original objects and a collection of current objects in hand. The collections are generic lists as in List<CustomerEntity>.

 

Step 1: Get a list of only those objects that have been altered since the data was first retrieved as well as objects to be inserted and objects to be deleted.

 

LINQ to the rescue!

 

In my example, origCustList is the List(Of Customer) with the original objects. custList is the List(Of Customer) with the new objects.

 

Inserts and deletes are not so complicated.

 

For inserts, I only need to query the current customers for those with no ID.

 

Dim inserts = custlist.Where(Function(c) (c.CustomerID=0))

 

Now I only need to add the new customers to my ObjectContext.

 

For Each insCust In inserts
   myObjContext.AddToCustomer(insCust)
Next

 

I cheated by using AddToCustomer for the inserts. For a more generic solution, just use the AddObject method instead.

 

 

For deletes I query the original customers and find any that don't have a match in the new customers.

 

Dim deletes = From oc In origCustList _
 
Where Not 
custlist.Any(Function(c) (c.CustomerID=oc.CustomerID))

 

Next you need to attach, then delete these objects from the ObjectContext. These steps for delete may not seem intuitive. The object needs to be known by the ObjectContext (therefore you attach it) and then deleting it from the ObjectContext marks it as a record to be deleted when we SaveChanges back to the database.

 

For Each delCust In deletes
   myObjContext.Attach(delCust)
   myObjContext.DeleteObject(delCust)
Next

 

Now come the updated cystomers. Here's where it gets fun!

 

I'm using LINQ to join on customerid and grab objects that have a newer dateModified value. This is where the date property pays off. Otherwise, I'd have to check every property of every object to see if anything changed.

 

Dim updates = From oc In origCustList Join cc In custlist _
On
oc.CustomerID Equals cc.CustomerID _
Where oc.ModifiedDate <> cc.ModifiedDate _
Select oc, cc

 

Now I have a pair of related objects for each modified customer.

 

In order to emulate a scenario where the objects have been tracked by the ObjectContext from the start, I need to attach the original objects to this new ObjectContext and then make any necessary changes to the objects. I iterate through the collection of object pairs and pass them off to a custom method, updateEntityProperty.


For Each c In updates
  myObjContext.Attach(c.oc)
  updateEntityProperties(aw, c.oc, c.cc)
Next

 

Here is where the cool stuff happens!

 

I want to preface this by saying that I tried using the ObjectStateEntry to pull this off, but it didn't give me quite enough capabilities as I needed for this solution. The class allows you to change the state of the entity to Modified and to change the state of the individual properties to modified, but I also needed to change the values of the properties, which I couldn't acheive with the state manager (which makes perfect sense - values are not state). [Update - apparently what I needed was in there; I just couldn't find it and after trying for some time, chose to move on to reflection.] So instead I am using reflection to help me modify the values programmatically. When I do this, the state changes automatically.

 

Although I didn't get to use ObjectStateEntry, I am leveraging the MetaDataWorkspace which is also part of the Entity Framework.

 

Private Sub updateEntityProperties(ByRef context As AdventureWorksModel.AWModel.Entities, ByRef origcust As Customer, ByRef newcust As Customer)


Dim mdw As Data.Metadata.Edm.MetadataWorkspace = context.MetadataWorkspace

'do an explicit declaration of Edm.EntityType, otherwise custtype becomes an edmtype
'which doesn't expose
properties that I want to grab through reflection!

Dim
custtype As Edm.EntityType = _
mdw.GetType(GetType(Customer).Name, "AWModel", Edm.DataSpace.CSpace)

For
Each p In newcust.GetType.GetProperties '<- that's Reflection at work
   Dim newprop = p.GetValue(newcust, Nothing)
   If newprop Is Nothing Then
     If p.GetValue(origcust, Nothing) IsNot Nothing Then
       p.SetValue(origcust, newprop, Nothing)
     End If
   Else
     If
Not newprop.Equals(p.GetValue(origcust, Nothing)) Then
      'if we are here, then the value in new cust is different from the original
      'SKIP THIS FOR NOW IF IT'S A PROPERTY COLLECTION -

      'TO DO: build in recursive behavior to update property collections as well
      'simplest test to see if it's a property collection or not: straight values are a System
       If newprop.GetType.Namespace = "System" Then
          p.SetValue(origcust, newprop, Nothing)
       End If
     End If
   End If
Next

End Sub

 

So, this is not quite perfect yet, but it does the job and I'm really happy that I got the concept worked out. I need to do a few more things to make this work generically and I also need to add in recursive mods to this. For example, the Customer has a property collection of addresses. In the above code, I am ignoring them, but I need to make this solution so that if it hits a property collection (this is where I'm testing for a System type - if it's not system type then it's a  property collection that is exposing child records) that it calls the same UpdateEntityProperties on that entity as well. SaveChanges will take care of all of the related records that are in the ObjectContext.

 

While this isn't a cut & paste solution like Mike's, i hope that the explanations help you to understand how and why this works. There are quite a lot of lessons that I learned during this exercise.

 

I have two more patterns that I have worked out which I want to write about. The first is dealing with the state when working in an ASP.NET application (postbacks trash the objectcontext, too) and the other is doing xml serialization on the entire object graph. Danny Simmons gave me a hint for how to approach the graph problem, and I was able to work it out from there.

Monday, October 01, 2007 12:09:34 PM (Eastern Standard Time, UTC-05:00)  #     |  Comments [2]  | 
 Thursday, September 27, 2007

We had a wee scare last night flying from Washington/Dulles into Burlington (on my way home from a quick trip to San Diego to do an INETA talk ... that blog post is still on my laptop :-)).

It was a late night flight, due to land at 11:30pm in Burlington.

Five miles before we got to Burlington, the airplane's monitoring system reported excessive heat in one of the front brakes. The worst case scenario was that on touchdown or cruising down the runway, the brake could catch fire and it was quite close to the engine so...scary potential.

The flight attendants (both young and neither had yet been initiated with an emergency landing before this) had to prepare us very quickly with the news and for an emergency landing. Tighten the seat-belts, tuck into a ball, brace yourself against the seat in front of you (I had a wall that was a little far away). There was no time really for much else. They did a fabulous job.

I live with a general disbelief in anything bad happening to me or my loved ones. My parents have both survived minor bouts with cancer; my father has been in two car accidents where the car was totalled and he walked away; my husband cracked his head open on a rock face while we were hiking and while there was a LOT of blood, missing teeth, a scalped scalp and huge gashes to be stapled up, he recovered perfectly and they even shoved his teeth back into his gums and they remain there today. My Newfoundlands are living to ripe old ages beyond the norm for this breed of dogs.

So yes, I live in a little fantasy bubble world.

And because of that, while I was certainly shaking a little, I didn't really think that anything bad would really happen. And it didn't.

We landed. Nothing seemed to happen out of the ordinary. We did an emergency evacuation of the plane because the potential for fire and explosion still remained. Since I was up front I just went down the stairs like I normally would (vs. jumping off the wing like a lot of other people had to). Nobody was hurt. The plane didn't catch fire.

There was a lot of excitement on the runway. Many firetrucks and emergency vehicles and two guys in Hazmat outfits who looked like astronauts inspecting the plane.

It struck me as odd that I was the only person who walked VERY far away from the plane (you know, just in case.). I just kept going further and further out in the field by the runway.  And an hour later, I was allowed to go back on the plane to get my stuff out of the storage bin and I headed home.

One poor girl had been so terrified (possibly she has some past experience that this too closely resembled) that she couldn't stop crying for at least an hour. She had her boyfriend there to comfort her.

So, I got another day out of it. Didn't even bother calling my husband since it was nearly midnight; I knew he was sleeping and all I had to report was "Hi honey. I'm here. The plane didn't crash or blow up and I'm still alive. Talk to you later."  But I call him with that report after every single flight, which means generally twice in a day, since most of my flights include one stop over. So it didn't seem necessary to wake him up for that.

Thursday, September 27, 2007 9:44:33 AM (Eastern Standard Time, UTC-05:00)  #     |  Comments [3]  | 
 Monday, September 24, 2007

One of the very difficult to solve problems in .NET client & web service solutions is sharing custom objects between the web service and the client.

Here's the scenario.

You have a business layer with some serializable objects. Let's say the namespace is "MyLovelyBAL" and the object is "MyLovelyObject".

You have an operation in your web service that returns one of those objects.

Now you build a client, say a Windows Forms app ("MyWinApp") that will call into that Web Service, and you are not savvy enough to use a tool like thinktecture's WSCF. :-) Therefore you "Add Web Reference" and point to the web service you built. In doing so, you are asked to give a name to the proxy. Let's call that "MyWSProxy".

In your WIndows Forms app, you also have a reference to the assembly that contains MyLovelyBAL.MyLovelyObject.

Now it's time to write some code.

You instantiate the proxy and request a lovelyobject.

Dim myproxy = New MyWSProxy.Service
Dim MyLovelyBAL.MyLovelyObject=myproxy.GetMyLovelyOBject

But it fails telling you that you can't cast a MyWinApp.MyWSProxy.MyLovelyObject to a MyLovelyBAL.MyLovelyObject.

And then you will begin on a journey to hell, which might include reading this article on Schema Providers which is not for the feint of heart.

So just forget it. It's 2007! We have WCF.

You can go into your project where you define the MyLovelyObject class and give it a DataContract attribute and give it's properties (that you want to expose) DataMember attributes. Your interface operation (and any operations that implement it) will return a MyLovelyBAL.MyLovelyObject object.

And when you Add Service Reference to your windows forms app, there is no namespace replacement. A "MyLovelyObject" is always coming from MyLovelyBAL and this is oh, so lovely. The way it oughtta be!

You know I learned this the hard way. And I will be seeing Michele Leroux Bustamante tomorrow who will first bop me on the head for waiting so long to use WCF and the sit me down and have me read her fabulous book Learning WCF, (which I had in my hands long enough to raffle off at VTdotNET last week BOO HOO) cover to cover. It's about time!

Monday, September 24, 2007 2:23:08 PM (Eastern Standard Time, UTC-05:00)  #     |  Comments [0]  | 
 Sunday, September 23, 2007

As we plug away on the prototype, the Astoria team is building brand new code for the real product and right now they are trying to finalize the URI query syntax and looking for your input. Read more....

[A New DevLife Post]

Sunday, September 23, 2007 7:21:55 PM (Eastern Standard Time, UTC-05:00)  #     |  Comments [0]  | 

I ran into what felt like a brick wall today trying to install Vista onto a VPC and finally found a way to accomplish it. Read more...

[A New DevLife Post]

Sunday, September 23, 2007 7:18:44 PM (Eastern Standard Time, UTC-05:00)  #     |  Comments [0]  | 
 Saturday, September 22, 2007

Boy am I excited about this.

Timothy Ng is on the VB team and wrote the recent MSDN Magazine article on Lambda Expressions for VB developers. I understood lambdas better than ever after reading the article ... finally something from the VB perspective.

His two talks are:

LINQ-ing your data
In this session, I will explore how LINQ will change the way you think about data in your applications. In particular, I will explore how LINQ provides a higher abstraction for accessing your data, provides a common abstraction across data domains, and allows for simplified data transformations. I will focus on LINQ to XML as the transformation technology, and the various XML features that are available in VB.

LINQ from the Ground Up
In this session, I will explore the fundamentals of LINQ by examining how the various language features (such as extension methods, type inference, lambda expressions, and anonymous types) form the foundation for LINQ. I will share some of the experiences that the VB compiler team had while building LINQ, in hopes that it will help you build your applications on LINQ. Examples will be in VB, but this talk applies to C# folks who are interested in learning how the fundamentals work as well. All language concepts apply to both VB and C#.

I sure wish I was seeing these before I do my LINQ talk in San Diego on Tuesday! :-)

Richard Hale Shaw will be doing some pretty cool LINQ talks too which will be a C# focus.

Before I learned that Tim was coming, I was already excited about the Code Camp speaker roster. This will be the first time that Jesse Liberty and Fritz Onion (who are "local" in about the same way that I am local... well, I guess Fritz's drive is even longer than mine). Jesse recently started working for Microsoft (see? I managed not to say "swallowed the red pill"!) and he is "da man" at Microsoft on Silverlight. Fritz is a true ASP.NET plumber who amazed us all with his book Essential ASP.NET in the .NET 1.0 days and again with the updated version for ASP.NET 2.0, because it explained HOW ASP.NET actually worked and he continues to spelunk on our behalves.

There are already almost 40 sessions listed with tons of amazing expertise on display.

I'll be doing a talk on Astoria, a session on using Entity Framework in multi-tiered apps (and getting a little nervous that nobody seems to be doing an Intro to entity framework session) and a session on ASP.NET Databinding with LINQ.

Saturday, September 22, 2007 9:33:52 PM (Eastern Standard Time, UTC-05:00)  #     |  Comments [0]  | 

I don't like not knowing how something works, so I just had to dig further into how Astoria services see requests for queries versus Updates, Deletes and Inserts.

In my previous post, I dug into the assembly with reflector and could see that a PUT was used for updates, DELETE for deletes and POST for inserts.

I turned tracing on for the Astoria service and pushed the trace info to a log file. You can see how to do that at the end of this post.

Looking at the messages, you can see the METHOD and the query operators but not the main call. :-(

For example if I made a call to service.svc/Customers the message looks like this. I've stripped out WebHeader information from the message.

<HttpRequest xmlns="http://schemas.microsoft.com/2004/06/ServiceModel/Management/MessageTrace">
<
Method>GET</Method>
<
QueryString></QueryString>
</HttpRequest>

I don't see that I called the Customers "operation".

If I request service.svc/Customers[52], it's the same. I don't see the Customers and I don't see the ID i've passed in.

If I use a query operator as in service.svc/Customers[City eq Seattle], it's still not exposed. Now I'm getting frustrated. I know I'll have to dig deeper at some point, or bribe Pablo to show me someday.

But for now here's a little gratification. If I create my own Service Operation which takes a query string, now there is something to see. My service operation is CustomersbyCity. the call looks like this

service.svc//CustomersByCity?city=London

The message looks like this

<HttpRequest xmlns="http://schemas.microsoft.com/2004/06/ServiceModel/Management/MessageTrace">
<
Method>GET</Method>
<
QueryString>city=London</QueryString>
</HttpRequest>

When I use my little console app as a client and do some inserting and deleting, I can see the methods coming up the pipe, remember, the Astoria.Client.dll translates the function (insert/update/delete) into a Method.

<HttpRequest xmlns="http://schemas.microsoft.com/2004/06/ServiceModel/Management/MessageTrace">
<
Method>DELETE</Method>
<
QueryString></QueryString>
</HttpRequest>

So, only MILDLY interesting to me (because I can't get in deep enough yet) and probably COMPLETELY uninteresting to most everyone else. But it was fun digging around and also learning how to use some of the WCF tool. Here by the way is how I got to see the messages.

Modify the web.config to "Turn On" tracing and messaging. You can do it manually or with a tool.

    <system.diagnostics>
        <sources>
            <source name="System.ServiceModel.MessageLogging">
                <listeners>
                    <add name="messages"
                    type="System.Diagnostics.XmlWriterTraceListener"
                    initializeData="c:\logs\messages.svclog" />
                </listeners>
            </source>
        </sources>
    </system.diagnostics>

    <system.serviceModel>
        <diagnostics>
            
<messageLogging logEntireMessage="true" logMessagesAtServiceLevel="true"
               logMessagesAtTransportLevel="true" />
        </diagnostics>
    </system.serviceModel>

 

You can use the WCF Config tool in Visual Studio 2008 to configure this. Right click on the web.config and choose Edit WCF Configuration. Here are screenshots of the settings that created the above in my web.config.

If you are using the Web Dev server, you need to stop that server so that the trace actually gets written to the file before you can look at it.

Once the file is created, you can look at it using the Service Trace Viewer which you can find in the SDK Tools (available from Start/Program Files/Microsoft Windows SDK for Visual Studio).

Saturday, September 22, 2007 3:54:12 PM (Eastern Standard Time, UTC-05:00)  #     |  Comments [0]  | 

I got bored with querying Astoria in the browser and and wanted to see how to use it in an application, so I opened up the sample console app that comes with the bits and started poking around and was astonished at what I saw.

You can create an new WebDataContext based on the Astoria data service.

WebDataContext is like an ObjectContext. Here's what it exposes:

namespace Microsoft.Astoria.Client
{
public class WebDataContext : IDisposable
{
public WebDataContext(string uri);

public string BaseUri { get; }
public ICredentials Credentials { get; set; }

public void Add(object obj, string entitySet);
public void Attach(object obj, string entityKey, ObjectState state);
public WebDataQuery<T> CreateQuery<T>(string queryString);
public WebDataQuery<T> CreateQuery<T>(string queryString, QueryOption queryOptions);
public void Delete(object obj);
public string[] GetEntitySets();
public void MarkModified(object obj);
public void SaveChanges();
}
}

So, you can create a WebDataContext

WebDataContext ctx = new WebDataContext("http://localhost:50000/northwind.svc");

then query it

Category beverages =
ctx.CreateQuery<
Category>("/Categories[CategoryName eq 'Beverages']"
)
.FirstOrDefault();

add a new object to the ObjectContext

Category newcategory = new Category();
newcategory.CategoryName =
"Sample"
;
ctx.Add(newcategory,
"Categories");

And then (this is the part that is sick and I want to rip open the covers and see how they are pulling this off!), call SaveChanges on the context!

ctx.SaveChanges();

Of course, you can update and delete, too.

So, I wanted to see what's going on here. It's REALLY important to understand that we're looking at prototype bits. THe production code that they are working on now will be different, though the concepts and use will be relative to what we can play with today.

You can't get any hints by adding ?WSDL to the end of the uri. Rembmer withi a default asotira service, there are no operations. It's all based on interpreting the URI. So I opened up Reflector and pointed to the key assembly.

SaveChanges calls a method called SendHttpEntityUpdate which sends up the object and the key and based on the operation calls a request method. So if you are adding, it sends a POST. If you are deleting, it sends DELETE and if you are modifying it sends PUT. The dll used by the service checks which method you pushed up (there's an enum for ReceiveEntityOperations) and processes the incoming data accordingly.

Neat!

So of course there will be the usual million concerns about this. There are definitely credentials built into these operations and all kinds of ways to limit what is available thorugh the web service.  And there's so much I don't know yet, so don't judge solely on what you are seeing here.

If you don't have a .NET Astoria client, you can just construct the POST, PUT and DELETE URIs. I can't say how to do that quite yet though. ;-) (Not that it's a secret. I just don't know how yet.)

But it's still darned cool. Nice job folks (and don't forget Alex who was key in getting Astoria going before he left Microsoft! )

Saturday, September 22, 2007 11:34:01 AM (Eastern Standard Time, UTC-05:00)  #     |  Comments [0]  | 

I find it helpful to use the browser to visualize immediate cause and effect of queries to an Astoria web data service. THere's so much more that you can do when you are really using this in applications, though.

In the documentation there's a list of query operators:

  • callback (specific to JSON for AJAX calls)
  • expand - pull related data using EDM associations
  • format - define format of the response (eg xml, json, rdf)
  • keyset- only return the entityKeys
  • orderby (you know what this is)
  • skip (same as in LINQ)
  • top

The doc explains each of these in more detail.

Here for example, I can get a specific customer and all of their sales order headers. (Make sure you have MultipleActiveResultSets=true in your connection string in order to do this. That allows the db to return more than one result on the same connection string .. a feature added in that we first acquired with .NET 2.0 and SS2005.)

I can drill even further to return the SalesOrderDetail data form the SalesOrderHeader using this URI:

http://localhost:55976/WebDataService1.svc/Customer[6]?$expand=SalesOrderHeader.SalesOrderDetail

Here's an example of using skip

Looking at the non-xml formats is tough in Internet Explorer, at least with JSON as IE doesn't know how to display the response. There are a few ways around this.

1) (bit of a hack :-))

Add this into the appSettings seciton of web.config

  <add key="EnableAstoriaJSONP" value="true"/>

and then browse your service in Firefox, Firefox will ask you waht you want to view the response in, choose notepad and you can see the json response. Read the caveats about JSONP in the Astoria documentation.

2) Use the helper tool which is an HTML page here [ http://astoria.sandbox.live.com/tools/raw.htm ]. You can't use this online (I know 'cause I tried in my impatience!). You need to download it and add it into your project. It will then properly parse and display results.

So all this is just about querying, the "R" (read) in CRUD. Wait till you see the C, U & D. Watch for another post.

Saturday, September 22, 2007 11:23:57 AM (Eastern Standard Time, UTC-05:00)  #     |  Comments [0]  | 
 Friday, September 21, 2007

Here's a quick start for checking out astoria. See my  blog post "what the heck is astoria" if necessary. :-)

Requirements:
VS2008 Beta 2
Entity Framework Beta 2
Entity Framework Tools August CTP
Astoria CTP September refresh (which works with VS2008 Beta 2)

Note that I had to do this in XP. I'm having a wierd problem in Vista around the ASPNETCompatibilityEnabled setting in web.config. You can follow the discussion of that if you are interested in this forum thread. The thread is accidentally a combination of Kevin Hoffman's issues with IIS and my issue with Cassini. They don't seem to be related problems.

1. Start a new web application project (not website) in VS2008.

2. Add an ADO.NET Entity Data Model using the projects Add New Item dialog.
If you haven't used the wizard before, you can see this blog post.
The basic steps are

  • Generate from Database
  • SElect a database (same as doing this for other data UIs in Visual Studio)
  • Just leave default names for things for this quickstart. (I pointed to Northwind in SQL Server and it defaulted at NorthwindModel and NorthwindEntities, etc)
  • Select all database objects (this is the default). This brings in tables, sprocs and views.
  • Then finish and you will see the EDMX file in the solution. You can double click on that to see the conceptual model (one of three parts of the model) in the designer. Note that this is just a simple way of getting a model from a database, the most basic type of conceptual model.

Add a Web Data Service.
Add New Item, choose Web Data Service. It's in the main section (Visual Basic/ C#, not categorized into Data or Web templates) down on the bottom, alphabetical. :-)

The services code view should open up. Note the TODO for the class. All you need to do is replace the stub [class name]

(VB)

(C#)

with the name of your entity wrapper class created in the model. THe model code gens a class file.
 

You can open up the model in teh designer and then see it's Namespace and Entity Container Name in the properties.
 

That's it! Now you can browse the web service directly (eg. right click on svc file in solution explorer and View in Browser) and start looking at how the data is exposed through the EDM.

My What the Heck is Astoria post walks through some quick query scenarios to give you an idea of what you can do.

This is just the tip of the iceberg of course. See the Astoria Team blog for more info.

Friday, September 21, 2007 10:36:45 AM (Eastern Standard Time, UTC-05:00)  #     |  Comments [4]  | 

I'll be doing a session on Astoria at REMIX Boston 07 (10/8-10/9) and at the New England Code Camp (9/29-9/30).

What is Astoria?

Astoria exposes data as web services through an Entity Framework Data Model. Once exposed, you can use URI based syntax to query data through the conceptual layer of the entity data model. You can control what is and isn't exposed and you can also do inserts, updates and deletes. The results can be output in a number of formats including JSON and ATOM.

It's simplest to understand this by looking at some queries in a web browser. The Astoria CTP add-in for Visual Studio 2008 Beta 2 has a template for creating a Web Data Service that looks like a WCF service. It only needs one class to expose the whole kit n' caboodle and that class inherits from WebDataService, a generic class. You merely need to tell that class that the type it is dealing with is the Entities class exposed by an Entity Data Model that you have created. The class doesn't even have any code in it.

  public class northwind : WebDataService<NorthwindModel.NorthwindEntities>
    {
    }

Then you can just make calls to that service and start playing with the query syntax.

The base service exposes the list of entities in the model (http://mysite/myservice.svc). The payload is just a raw XML representation of the data. There will be payloads for different formats, such as JSON, ATOM and more. See this Astoria team blog post for more info.

Then you can start drilling in. Here for example, I have asked for the Categories entities. It will return all entities with their details.

Now you can start seeing the URIs that are displayed to show you how to drill further into data. When looking at it in the browser, you might want to click on those URIs, but remember, in a real app this is not how you will be working with the results of the service call. :-) Note that you can see how to get to the related Products data although it is not pulled in automatically. That's no different than how Entity Framework works by default - deferred loading of related data.

So let's try the Category uri for categoryID 2.

And then get the products for that category.

So you can see now the basic concept of what Astoria is doing. I am not forced to write a billion operations in my web service to expose my data every which way. I don't have to deal with building xmldocuments to return data that can be consumed by anyone, not just .NET clients that can deal with serialized datasets. With proper security in place and the possibility of customizing what is and isn't exposed (and how), consumers of my data can now have easy access to do all kinds of reporting, mash-ups and more. The fact that any consumer will know what to do with any data that is exposed via an astoria web data service is huge.

You can learn more about Astoria by following the Astoria Team blog.

Friday, September 21, 2007 9:14:28 AM (Eastern Standard Time, UTC-05:00)  #     |  Comments [0]  | 
 Wednesday, September 19, 2007

When Vermont.NET does not have a sponsor to cover the cost of pizza & soda, the attendees pay for their own - $5 each. This was the case at Monday's meeting. Additionally, we were doing a fundrasing raffle for the first time ever, using a license to Infragistics NetAdvantage that was donated by Infragistics. The raffle was $5. There were about 25 people at the meeting.

Usually, collecting the money is a subtle thing that happens in the background and I have not even had to do it in years as another user group member has been grabbing the cash on my behalf. But I was doing it on Monday and wouldn't you know it, but everyone had 20's. Nobody seemed to have change. In over 5 years and 70 meetings, this has never happened before. It took a good 15 minutes or more to get it all sorted out. While I'm laughing about it now, it was totally chatoic and embarrassing at the time. We had a bunch of new attendees and our speaker, Mark Mullin (who helped come up with change for a 20) from New Hampshire. I could only imagine what they were thinking!

So my lesson? I guess from now on, I'll just have to add in "20's not accepted".

 

Wednesday, September 19, 2007 9:17:06 AM (Eastern Standard Time, UTC-05:00)  #     |  Comments [3]  | 
 Tuesday, September 18, 2007

Erick Thompson is at SQL PASS and will be doing a talk on Entity Framework for DBAs tomorrow (Wednesday) afternoon.

You can get a preview of what he will be talking about by reading his recent post on the topic.

Tuesday, September 18, 2007 7:35:55 PM (Eastern Standard Time, UTC-05:00)  #     |  Comments [0]  | 

While there are many great LINQ resources for VB, I happened to read/watch/see these and wanted to be sure to point them out.

1) Lisa Feigenbaum's LINQ Best Practices webcast. Lisa is on the VB team. This webcast is full of great insights and tips.

2) Timothy Ng's Basic Instincts: Lambda Expressions article in the current (Sept 07) issue of MSDN Magazine. I've spent so much time trying to interpret lambda articles written in C#. Having one written specifically for VB devs is awesome and this article gently leads you into the complexities in a very easy to comprehend way.

3) 101 LINQ Samples for VB (yay!) was posted a few weeks ago. LINQ to Objects, to SQL, to XML and to DataSets. Don't go looking for LINQ to Entities yet though.

 

VB
Tuesday, September 18, 2007 12:24:42 PM (Eastern Standard Time, UTC-05:00)  #     |  Comments [0]  | 

Inc's 5000 list was recently released and among the 9 Vermont companies on the list there are two software companies, both based in South Burlington.

 iTechUS ranked 113 on the whole list and #11 in IT companies. In New England it is ranked #1 in the IT Services companies. www.itechus.com

Dealer.com ranked 578 in the full list. While they are a software company, their industry is listed as advertising and marketing. In that category they are #55. www.dealer.com

Tuesday, September 18, 2007 10:04:51 AM (Eastern Standard Time, UTC-05:00)  #     |  Comments [0]  | 

Just a quick pointer to Pablo Castro's announcement that a new CTP for Astoria is out that will work with vS2008 Beta2.

An important note, this is a refresh of the CTP, which is based on the early prototype. The actual production code that they are working on is a totally different set of bits. Hmm, for safetey's sake, I better just quote Pablo:

NOTE: let me stress that this is a refresh of the May CTP, which is based on the initial prototype of Astoria. All of the features/design options we’re discussing here in this blog (e.g. last post about payload formats) are in the context of the production version of Astoria which we are in the process of building. The production version incorporates a ton of feedback that we received over the last few months plus our own more detailed thinking of what should the system look like. So expect to see quite a few differences between the CTP and the production code once we start shipping the production version. The principles will all be the same, but the details will vary.

Tuesday, September 18, 2007 6:05:17 AM (Eastern Standard Time, UTC-05:00)  #     |  Comments [0]  | 
 Monday, September 17, 2007

This a job with one of my mentoring clients, Co-Operative Insurance. I absolutely love going to their office and working with these people. It's a small team in a large company. Nice, friendly folks. And it's in Middlebury which is a great place to work.

Web Application Development

Excellent opportunity for a qualified programmer with web application design, development, and implementation experience!

Work on the complete product development lifecycle of our B2B extranet, including requirements gathering, technical design specification creation, coding, testing, deployment, post-production support, documentation and maintenance.

Must have 2-5 years’ development experience with Visual Studio 2003/2005 as well as JavaScript, (X)HTML, ASP.net, VB.net and XML. Pluses include insurance industry knowledge and experience with CSS and with version control software such as SourceSafe or Vault. Solid UI design skills get you multiple pluses.

To join our hard-working, fun IS team in producing quality apps, you must be an organized, self-motivating team player, an excellent communicator and attentive to detail. You must be willing and able to work both independently and with a team, depending on the assignment.

Co-operative Insurance Companies is a $50 million property and casualty insurer in VT and NH.  We offer a competitive benefits package and a schedule that allows for work/life balance.

For immediate consideration, please forward a copy of your resume to jobs@ciui.net.

Monday, September 17, 2007 12:20:58 PM (Eastern Standard Time, UTC-05:00)  #     |  Comments [0]  | 

We are really looking forward to this talk by Mark Mullin tonight at VTdotNET!

We also have some great WCF books to give away: Learning WCF by Michele Leroux Bustamante (thanks OReilly and INETA) and Pro WCF: Practical Microsoft SOA Implementation from APress.

Thanks to Infragistics, we will be raffling off two licenses to NetAdvantage for .NET. One of the raffles will actually be a fund-raising raffle.

We also have shirts and swag courtesy of CodeZone and piles of CoDe Magazine and aspnetPRO to give away.

And thanks to Verio's sponsorship of INETA, I am able to offer Verio's free webhosting to Vermont .NET members who show up at the meeting tonight.

Since it's such a nice day (and I have just absolutely NO work to do ;-)), and Mark is being so generous with his time (driving up from New Hampshire) and cash (there's gas $ and hotel involved), I am going to take him on a 1.5 hour cruise on Lake Champlain aboard the Spirit of the Ethan Allen this afternoon. Hey, running a user group is HARD work, y'know?

Monday, September 17, 2007 11:15:14 AM (Eastern Standard Time, UTC-05:00)  #     |  Comments [0]  | 
 Sunday, September 16, 2007

Loren Heiny has been working on a cool Silverlight annotation app to user Silverlight and a stylus to do Google and Technorati searches. It's called SearchTip and is live on TabletPC Post.

There is a lot that is really interesting about this. (And it's visually appealing too!) First of all, he has a YouTube video where he compares using this (earlier version) on a Mac to using Inkwell on the Mac (when you have a tablet ala Wacom attached to the Mac) and demonstrates the superiority of Microsoft's handwriting recognition. I was lucky enough to attend an SDR at Microsoft a few years ago where the man who ran the whole reco team explained to us how they made it all work. I can't imagine anybody being able to come close to their investment or acheivement with handwriting recognition, so it was great to see Loren being able to really compare apples to apples (no pun originally intended, but oh well!)

Next is the fact that Loren is doing handwriting recognition from the InkPresenter in Silverlight. Not an easy feat. Once the reco is done, the result can get passed to a google service and voila you get your results displayed.

Loren has done some amazing work with the Tablet APIs for a number of years and always thinks out of the box. So I love that he's taking his ideas and applying them to Silverlight.

 

Of course, you don't need a tabletpc to draw in silverlight, but as awesome as the recognition is, I don't think it's fair to ask it to recognize mouse drawn handwriting. But it's not bad. Reco returns a list of guesses, sorted by best to worst. Here I'm using an earlier version which does google searching and trying it with my mouse on a regular (non-tablet) desktop.

So you can click on the guess (I clicked on Steins) and it will show you the next guess on it's list. Heinz was next, then a few more clicks (just to see) got me off the track. oops. This is not a reflection of the reco or of Loren's app. I'm only playing around with the mouse for fun (and laziness since my tablet is upstairs.) So while the reco may not be great when you aren't using a stylus which has about 10 times the resolution as a mouse, Silverlight's ability to collect the ink data from the mouse is still pretty impressive.

You can see Loren doing this. and how impressive it truly is using a Tablet, in his YouTube demo video.

You can clear your writing with the x, and of course, when you have the correct reco displayed, click Search.

Sunday, September 16, 2007 9:48:13 PM (Eastern Standard Time, UTC-05:00)  #     |  Comments [1]  |