Querying Entity Framework. Part 23 – Filtering Nullable Field

When you run queries against non-nullable columns (NOT NULL in SQL), everything is going smoothly. You should pay some attention when you deal with nullable fields (one of the cases will be covered at the end of this post).

Now let’s look closer to Entity Framework queries against nullable field. I’ve added a new column LastUpdate with the type DateTime? to my Client entity.

So I have some records with concrete dates, and some with NULL values.

1. Filtering “greater or equal than”

C# understands it perfectly. We have a simple and correct SQL query.

2. Filtering all concrete dates aka WHERE … IS NOT NULL
You can use one of the choices.

Both generate the same correct SQL query.

3. Filtering multiple dates aka WHERE … IN

This case is a bit harder because C# could not compare DateTime[] array and DateTime? field. You need to get a not-nullable value from column.

Looks good, but I’m not satisfied with conversion of C# DateTime array values to SQL Server DateTime2 type. Especially when I set a strong annotation to use Date type in database.

If you don’t want to build a query with every parameter despite whether it is null or not, you might build a dynamic filter like in the following sample.

What will be if both parameters are null? No filters will be added, and you’ll get a query returning all(!) rows regardless of having a concrete value or not. In this case add another check when both parameters are null. One of the simplest implementation is to add third if statement.

Unit Testing C# Async Methods

This time I have a C# async method that should be tested. As you remember, async method must return Task or Task<>.  You can declare an async void method, but this should be used only for event handlers because you have no control on method execution and, the most important, its failure. Beside that, async void method is hard to test.

I have a simple async method that returns Task<int> instead of int in synchronous method (look at my previous post Unit Testing C# Synchronous Methods).

We must be happy that modern unit test frameworks allow to write async unit test where an async method is called asynchronously.

1. Check successful result

The async method being tested is called via await operator. This makes the code being executed in true asynchronous mode.

2.1. Check failure with ThrowsAsync<>

Unit test is awaiting for result from Assert.ThrowsAsync, that is awaiting for result from the method being tested. If you delete the inner async/await, the method would be executed in synchronous mode. If you omit the first outer await, the unit test method might finish before the code in NumberAsync would fail. So you will get wrong results!

2.2. Check failure with Record.ExceptionAsync

Unit Testing C# Synchronous Methods

When you need to unit test a method, you should check happy path (for example, the method returns a resulting value) and sad path (the method throws exception). Here I describe a basic usage of unit testing with Visual Studio 2017 and xUnit version 2.2.

The code being tested:

1. Check successful result (happy path) – it’s very straightforward.

2. Check fail when the method throws an exception (sad path)

If you have faced with MSTest, you might remember [ExpectedException] attribute. In this case MSTest waits for a particular exception would be thrown in a whole unit test method, but not in a specific line of code. Modern unit test frameworks have more graceful capabilities to catch the exception.

2.1. Using Throws<>

But this approach combines Act and Assert phases of unit test in one line of code. Richard Banks suggested a better way in his article Stop Using Assert.Throws in Your BDD Unit Tests.

2.2. Using Record.Exception

At first, I check that the exception was really caught, then check the type of that exception.

Next time I’ll tell about unit testing the asynchronous methods.

ASP.NET Identity. Part 2 – Customizing Entities

ASP.NET Identity gives a class ApplicationUser for storing data about users. It has a lot of properties defined in base class IdentityUser, such as Id, UserName, Email, PhoneNumber, etc. But when you need to add some extra properties, it becomes more complicated.

Entity Framework creates a database only once at first use. It does not recreate, nor modify it. We should use EF migrations to add new fields. Also it gives the possibility to change table names.

1. Add new properties in file Models\IdentityModels.cs

2. Open Package Manager Console and run command

EF will create folder Migrations and a couple of files: Configuration.cs and <timestamp>_InitialCreate.cs with code to create a default database structure.
3. Create new migration and give it a name

Visual Studio will open a file with two method to add and drop columns defined in step 1.
4. Apply EF migration to database.

Now I have the following database structure:

5. I want to enter first and last names during the registration process. So I’ll add two properties to class RegisterViewModel in file Models\AccountViewModels.cs

6. Add code to Views\Account\Register.cshtml file to display these them.

7. Change Register method in Account controller (file Controllers\AccountController.cs) to store them.

8. Also I want to display user’s first name on the top bar of the web application instead of his/her email. To do this we need to modify file Views\Shared\_LoginPartial.cshtml to get ApplicationUser via current user’s Id with the help of Owin context, then to show property FirstName. The final code looks like this:

ASP.NET Identity. Part 1 – Changing Default Database

When you create MVC project in Visual Studio 2017, you can use individual user accounts authentication that is shipped with MVC template. This implementation uses Entity Framework. By default, EF creates a database in SQL Server LocalDB with a name like aspnet-WebApp-<timestamp>.

To change the connection to database:
1. Open Web.config file
2. Find section <connectionStrings>
3. Modify the attribute connectionString in element with name DefaultConnection. Element will look like the following:

<add name="DefaultConnection" connectionString="Data Source=(local);Integrated Security=SSPI;Initial Catalog=AspNetIdentityDB;" providerName="System.Data.SqlClient"/>

If you dislike the name DefaultConnection, you should change it in Web.config file and in the constructor of ApplicationDbContext class (file Models\IdentityModels.cs).

Creating GitHub Project In Visual Studio 2017

My goal is to create a new solution and place it in GitHub. This can be done in few steps:

1. Create a GitHub account.
2. Install a GitHub extension for Visual Studio. I’ve downloaded it from https://visualstudio.github.com/
3. Start Visual Studio, open Team Explorer and connect to GitHub.

4. Create a new GitHub repository.

Click link, and enter data in the next form.

5. Switch back to Team Explorer. Click Create a new project or solution link.

6. Choose project type, give it a name. Step forward on wizard steps.

After that you will have the following structure in Visual Studio

and in a disk folder

Now you can write code, commit changes and push them to GitHub.

Converting List To IDataReader

I wrote about using SqlBulkCopy to fast load data from .NET into SQL Server database. This class needs a DataTable or IDataReader instance as a source. You can convert a C# List to DataTable (look here Converting List To DataTable). Now I’ll show a couple of examples how to convert List to IDataReader.

1. Straight implementation of IDataReader interface
I’ve seen a couple of good examples from Bruce Dunwiddie https://www.csvreader.com/posts/GenericListDataReader.cs and Venu Gopal http://technico.qnownow.com/custom-data-reader-to-bulk-copy-data-from-object-collection-to-sql-server/. Based on these two solutions I’ve made my one.

Now you can iterate through IDataReader like you do with SqlDataReader.

2. FastMember NuGet package

Thank you, Marc! it’s a good job https://github.com/mgravell/fast-member

You need to install NuGet package FastMember, add using FastMember; statement, and run this code.

Finding All Controls In ASP.NET Page

When you work with ASP.NET page, you might need to enumerate all controls of a specific type, i.e. text boxes, dropdown list, or table cells. All these controls derive from System.Web.UI.Control class.
To find all controls you need to iterate them recursively from the top one, i.e. Page, or any element that is located on a page (panel, table, etc.). One of the solution is to write an extension method like this.

Also you can use a generic version.

When I need to find all table cells with a specific ID, I can write the following LINQ query:

In C# 6.0 and later you can use Elvis operator in Where clause.

Converting List To DataTable

In my previous post SqlBulkCopy I wrote about how to load data from .NET code into SQL Server database using efficient BULK INSERT command. This class needs a DataTable or IDataReader instance as a source. I’ve collected a couple of examples how to convert List to DataTable.

1. Variation of Marc Gravell solution.
Source: https://stackoverflow.com/a/14548027

2. Solution created by Jennifer Hubbard, Bill Wagner, etc.
Source: How to: Implement CopyToDataTable<T> Where the Generic Type T Is Not a DataRow

SqlBulkCopy

SQL Server have great commands like BULK INSERT or OPENROWSET(BULK…) to insert a huge amount of data into a database. .NET Framework has a class SqlBulkCopy that provides similar capabilities in C#. Full documentation is available on MSDN article SqlBulkCopy Class. But SqlBulkCopy has some restrictions on input data. It can read from:

  • DataRow[] array;
  • DataTable;
  • IDataReader.

A simple example of usage can be as follows:

SQL Server Profiler shows that the correct command was really executed.

In order to use SqlBulkCopy class elegantly, it would better to have a method or extension to convert a generic list to DataTable or IDataReader. But it’s a topic for my next post.