Entity Framework Migrations and Data Seeding

This article is a part of a multi-post article on Entity Framework Code First for web applications:

1. Entity Framework Code First Entities and entity relationships
2. What do you need to know about Entity Framework Data Context
3. Entity Framework Migrations and Data seeding

This article is the second part of the posts. The first part talks about defining entities and relationships. The second part talks about the data context. You may find links to previous posts above. Let’s continue by talking about how to propagate changes to the database and seed the database with data for referential tables.

Entity Framework Migrations is a feature of Entity Framework to manage the reflection of changes in the code to the database structure. It compares the structure of entity classes and database tables to identify changes and write SQL scripts for implementing changes in the database.

To start using Migrations feature, we must first activate the tracking of changes in entity classes. Enabling Migrations can be done with a simple command which can be given through Package Manager Console (you may find it at: View->Other windows->Package Manager Console). In the console, you write the command Enable-Migrations (please note, this command may be different in versions of Entity Framework, my examples are using Entity Framework 6).

Before executing the command, please note to have the right project selected as the default project (in case you have an n tier architecture with data access layer on it’s own project).

After the command is executed, you will see a new folder appearing in the project structure, which holds the migration related files. As you add migration points, you will see files to be added to this folder with the name of migration point and its timestamp.

Inside the Migrations folder, you will also find a file named Configuration.cs. This file contains the initial configuration for the database, which we may use to populate some database tables with initial values and many other configuration tasks. A sample initialization can be done through the Seed method of the Configuration class:

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

internal sealed class Configuration : DbMigrationsConfiguration<MyFinancesContext>     {         public Configuration()         {             AutomaticMigrationsEnabled = false;         }           protected override void Seed(MyFinancesContext context)         {             var cashPaymentType = new PaymentType() {Active = true, Name = "Cash"};             var visaCardPaymentType = new PaymentType() {Active = true, Name = "Visa Card"};               context.PaymentTypes.AddOrUpdate(cashPaymentType);             context.PaymentTypes.AddOrUpdate(visaCardPaymentType);               var foodCategory = new Category() {Active = true, Name = "Food"};             var rentCategory = new Category() {Active = true, Name = "Rent"};               context.Categories.AddOrUpdate(foodCategory);             context.Categories.AddOrUpdate(rentCategory);               context.SaveChanges();         }     }

In the example above, you may see two sample classes, PaymentType and Category, which refer to referential tables in database, being populated with initial values, and the context saving changes at the end. This happens every time we issue update to the database. To avoid duplicate entries, please use AddOrUpdate method to add entities, instead of just Add.

But, what do I mean by a migration point. As you develop, often you make several structure changes in the classes, you add new classes, you add properties to existing classes, refactor existing ones, etc. After you do a fix or complete e feature, you create a migration point which stores changes from previous migration point until now. We can create a migration point by executing this command:

PM>add-migration 'name of migration'

Each successful migration point will create a file in the Migrations folder named as timestamp_name_of_migration.cs. These points will help you if you want to go back to a previous state of the database structure, in case you want to roll back the developments.

One point to be clear, when you create a migration point, the changes are not yet pushed to the database. It is only when we execute command Update-Database that the Entity Framework tries to push changes to the database based on the connection string found in the config file. This MSDN article shows a more complete list of migration commands.

We might some times want to update the database ourselves and not let the Entity Framework do it on its own. We may ask Entity Framework to show us the script of changes. We can do this using the command Update-Database -Script. After the execution, we will have an SQL script with all changes pending to be executed in the database.

Summary

Entity Framework Migrations is an important feature which will allow you to adopt the database structure in agility as you develop the code. Changes you make are kept in history and if you would like, you may go back in history, track changes, or even revert back to some point in history. All this with an easy command line interface which is easily executed in the Package Manager Console. Find more info in the official documentation of Entity Framework.