EFCore.Seeder 2.0.1

Due to nuget.org formatting issues, some code sections below are incorrect. You can find the original version of the documentation here

EFCore.Seeder

A library to seed databases from CSV files, using .NET Core 2.0 and Entity Framework Core 2.2.1

A project based on dpaquette/EntityFramework.Seeder, using EntityFramework Core instead of EF6, plus some improvements on how to handle resource files.

How to install

Install-Package EFCore.Seeder

How to use

TL;DR;

1. Configure Entity Framework Core in your solution and get a DbContext ready.
2. Add CSV files to your solution and set Build Action on each one of them to Embedded Resource
3. Make sure each one of the entities you want to insert or update implement the IEquatable<Class> interface. Also, if your entities include an Identity column, make sure to add the attribute [Ignore] to those, or make sure to set IDENTITY_INSERT ON on those tables where you want the tool to insert the identity columns you include in your CSV files.
4. Create a new CsvHelper.Configuration.Configuration that conforms to your CSV files. Please refer to their website for more information.
5. Create a ManifestConfiguration so EFCore.Seeder can find your CSV files.
6. Add this line: SeederConfiguration.ResetConfiguration(csvConfiguration, manifestConfiguration, assembly: typeof(<Assembly>).GetTypeInfo().Assembly); where <Assembly> is the name of one of the classes in the same assembly where the CSV files are included. csvConfiguration and manifestConfiguration are the previously created instances.
7. Add this line (and repeat as necessary): <DbContext>.<DbSet>.SeedDbSetIfEmpty(nameof(<Resource>));. SeedDbSetIfEmpty is an extension method on DbSet, and <Resource> is the name of the CSV file you want to seed into the database (without the extension ".csv")

Configuration and setup

This library requires all resource files (CSV files) to be added as Embedded Resource in a runtime available assembly. Once that requirement is met, the seeder needs to be configured using:

SeederConfiguration.ResetConfiguration(assembly: ResourceAssembly);

In order to make it easier, you can reference a type in that assembly and let Reflection take care of retrieving the assembly instance

SeederConfiguration.ResetConfiguration(assembly: typeof(<Assembly Class or Type>).GetTypeInfo().Assembly);

Also, if the format for the CSV files needs to be changed, or the CsvHelper configuration needs some tweaking, this is where we need to pass in the new parameters:

SeederConfiguration.ResetConfiguration(csvConfiguration: csvConfiguration, manifestConfiguration: manifestConfiguration, assembly: typeof(PayrollContext).GetTypeInfo().Assembly);

Where csvConfiguration is an instance of CsvHelper.Configuration.Configuration, from the CsvHelper library, and manifestConfiguration is an instance of ManifestConfiguration. The last one is just a way to format the embedded resource file names to make them easy to find in the assembly. You can store the configuration in a json file and load it when setting everything up.

"manifestConfiguration": {
        "delimiterFieldName": "{delimiter}",
        "resourceFieldName": "{resource}",
        "extensionFieldName": "{extension}",
        "format": "{delimiter}{resource}{delimiter}{extension}",
        "delimiter": ".",
        "extension": "csv"
    }

As written in that example, the library would search for a resource file, in the assembly we indicate, which would be formatted as follows

.<DbSet Entity>.csv

Then the library would get the first ocurrence of a resource which name is as indicated. If we point to a Products manifest, we'd need to name our file Products.csv. Doesn't matter which folder we put it in, or how deep in the hierarchy goes. As all embedded resources would be referenced as <Assembly><Path to Resource><Resource filename>, it would be picked up without problem.

With regards to CsvHelper configuration, please refer to their website.

Usage

Once the Seeder is configured, all we need to do is call the appropiate extension method when accessing a DbSet for a particular type.

Let's say we have a DbContext with a Products DbSet. We could do the following:

dbContext.Products.SeedDbSetIfEmpty(nameof(context.Products));

This would assume that we have a Products.csv file in the configured assembly, with all the required information to load into the Products entity.

Also, please note that if we're going to update information as well as insert it, the Products entity must implement the IEquatable<T> interface, so we can use product.Equal(other) when finding the right entity to update. This is due to Entity Framework Core not having an AddOrUpdate method, so we need to either use the method Add or the method Update.

Please check out the test projects for more information on how to use the library.