Skip to content

How to seed your EF Core database

A typical requirement when developing new applications is to have the ability to populate your database with an initial set of data. This may be to enable a common set of lookup data, or even to enable populating it with an inital set of data.

We may even want to populate your database with some Mock testing data, in order not too expose actual customer data to your developers or even to leak potentially sensitive data.

Using Entity Framework Core as an Object Relation Mapper (ORM) for your Core Web or API projects, is really easy and you can be up and running with it really quickly. In this post I will discuss how to seed your EF Core Model first database with seed data by using migrations.

I have previously discussed how to use EF Core in a seprate class library project and how to configure migrations to enable the update of your database schema.

Often you will also need to Seed your database with an initial set of data.

Database seeding is the initial seeding of a database with data. Seeding a database is a process in which an initial set of data is provided to a database when it is being installed . It's a process which is typically only run once on the initial deployment of a database.

I generally always follow principles from Code Complete:  A Practical Handbook of Software Construction when designing my data tables and schema's - Chapter 18 is definitely worth a read - the most common mistake I have found when working on legacy applications, is that the developers have often attempted to design their databases for all possible use cases, i.e. Application, Reporting, Data Warehouse, transactional.

This often results in bloated databases and applications which eventually become increasingly difficult to manage, maintain and develop against. This is one of the driving reasons why micro-service architectures have become popular, whereby the application and database are just enough to solve the problem.

You'll notice that I have tried to keep the database behind this service really small, and only implemented the Minimum Viable Product (MVP) requirements. We only need the ability to Add, Edit and View a list of Web Crawlers and their referer URL's.

Update .net core 2.1

Entity Framework Core 2.1 added new seeding features. Contrary to the older version of Entity Framework database seeding now works together with Migrations.
I have left the older version of this blog post in for reference and have marked it as legacy implementation should others be referring back to it.

Seeding with OnModelCreated

It is possible to initially seed your database directly after it has been initially been created. This can easily be accomplished by utilising the HasData method.

The HasData is defined with the EntityTypeBuilder which is returned from the Entity method of the ModelBuilder. It has been defined with a params keyword which enables any number of entity types to be passed in.

Lets look at an example, where we could use the HasData method. One example could be where we are defining a really simple Entity. A point to note is that it has Fluent API, which you use to apply in the OnModelCreating method.

Our Skill Entity is a simple Lookup type entity, which will be foreign key to a CandidateSkill table

public class Skill
  public Skill()
     CandidateSkills = new HashSet<CandidateSkills>();
   public int Id { get;set; }
   public string Name { get; set; }

We can now create our Candidate context and create the table

public class CandidateContext : DbContext
   public CandidateContext()
   public CandidateContext(DbContextOptions<CandidateContext> options): base(options)
   public virtual DbSet<Skills> Skills { get; set; }
   protected override void OnModelCreating(ModelBuilder modelBuilder)
     modelBuilder.Entity<Skills>(entity =>
      entity.HasIndex(e => e.Skill)
     entity.Property(e => e.Id).HasColumnName("id");
     entity.Property(e => e.Skill)
             new Skills {Skill = "C#",  Id = 1},
             new Skills {Skill = "PHP",  Id = 2},
             new Skills {Skill = "HTML", Id = 3},
             new Skills {Skill = "JavaScript", Id = 4}

You'll notice using this method we explicitly set the key property, Id, with a value.

With this in place we can generate a Migration, in this instance it could be the Initial Migration dotnet ef migrations add initial

Inspecting the migration file, that is generated for this, you see it generates the following logic


HasData is specific to a single entity. You can’t combine inserts to multiple tables with HasData.

If you use the dotnet ef migrations script command to check the generated SQL script EF core will use insert the data, you'll notice that depending on the Provider EF core will generate the appropriate SQL.

In this case because we generated the script to user MS SQL, and we elected to hard code id EF core will wrap our Insert statements with IDENTITY_INSERT


The Database

The database we are going to seed in this example is a relatively simple database with just enough complexity in it to illustrate how you can cope with most database seeding tasks.

We have 3 tables Threat, ThreatType and Status. Each table has a Primary Key defined and the Threat table has foreign key relationships with both the ThreatType and Status tables.

We are going to seed all 3 of these tables with data, including the relationships.


From this point on is the Legacy Approach


Although we have created a separate class library project for our EF Core database context class, ApiContext, we are going to develop some extension methods to this class but we are going to add this extension class to our main Api project.

The primary reason being that our Api.Database project is generic mini database project that is going to used for several disconnected micro service applications internally. Each of these micro service applications may have different database seeding requirements and therefore we want to be able to control the seeding on a per project basis.

Add Extension class

Add a new class to the root of the Api project and name it DbContextExtension.cs and add the following code.


View Full Code

We have added a simple boolean method we will call to determine whether all migrations have been applied.

Add Folder for data seeding scripts

Add a folder that will contain your data seeding scripts. This fodler will contain some JSON files which will contain the data required by the applicaiton.

I have just added a folder and called it seed. I have added three JSON files, threats.json , status.json & types.json

Add Seeding Extension Methods

We need to import a few additional namespaces to the DbContextExtension.cs


Add and addtional extension method to the DbContextExtension.csclass, that will be responsible for reading the data from the JSON files and the inserting to the database.


View Full Code

Modify StartUp.cs

We now need to add some code to the StartUp.cs to the Configure method



The above techniques illustrate how it easy it is to make use of Entity Framework core to create a database and seed it with the default data that it requires.

We used standard JSON data as a source for our data. We also made use of EF Core functions to insert the data.

Gary Woodfine
Latest posts by Gary Woodfine (see all)