How to configure an Identity column using Entity Framework Core?

Question

How do I create an Auto increment identity column in Entity Framework Core?

Obviously I can do it using fluent API for EF6 for example.

Answer 1

Looks like they changed it once again and the new method now is:

.UseIdentityColumn()

Answer 2

With latest bits EF Core 1.0 and up you should use

builder.Entity<ApplicationUser>().Property<int>(nameof(ApplicationUser.AccountNo))
            .UseSqlServerIdentityColumn()

Answer 3

In latest version of EF7 there is a new extension method to set identity column

protected override void OnModelCreating(ModelBuilder modelBuilder)
{    
  modelBuilder.Entity<MyEntity>(b =>
  {
    b.HasKey(e => e.Identifier);
    b.Property(e => e.Identifier).ValueGeneratedOnAdd();
  });
}

Answer 4

Here is how to do it explicitly in the event that you want to OR it's not the default behaviour.

protected override void OnModelCreating(ModelBuilder modelBuilder)
{    
  modelBuilder.Entity<MyEntity>(b =>
  {
    b.Key(e => e.Identifier);
    b.Property(e => e.Identifier).ForSqlServer().UseIdentity();
  }
}

Answer 5

Since there is very little EF7 documentation, much of what we know we have to glean from the source or unit tests. According to the following two unit tests in the EF7 source...

Here and Here

You would configure a property for Identity like this:

b.Property(e => e.Id).ForSqlServer().UseIdentity();

And you would configure a property for Sequences like this:

ForSqlServer().UseSequence();

The urls have changed due to the aspnet-core reorg, and the methods have also changed since this was first asked.

Here and Here

if (_useSequence) 
{
    b.Property(e => e.Identifier).ForSqlServerUseSequenceHiLo();
} 
else 
{
    b.Property(e => e.Identifier).UseSqlServerIdentityColumn();
}

It's possible these urls might change again (which is why I include the relevant code), but it's ridiculously easy to just look at the url and go to the site and figure out what the new url is.

Really, the whole point of my answer is that you can figure this stuff out yourself just by going and looking at the unit tests in the source code on GitHub. You shouldn't need someone to spoon feed it to you.

EDIT: Updated links to version 2.1 (still works for 1.1 and 2.0 as well)