Getting Started with Marten
As I wrote in my previous blog post, PostgreSQL is a very capable SQL database. But one thing very interesting about PostgreSQL is its excellent support for JSON documents.
The ability to store JSON documents in a database basically allows the database to run like a NoSQL document store, effectively combining relational and non-relational into one persistence option.
Combining NoSQL and SQL into the same persistence store is a common goal. The reason is we often need to store both types of data. We often have a lot of very rigid data that fits great into a SQL option, but we often have some data that is too variable to store in a SQL or relational structure.
I have worked on teams where we have created crazy DB structures to handle the storing of data that didn’t fit well into a relational structure.
Storing that non-relational data into a JSON or an XML column makes a lot of sense; many of us have tried to do it. Unfortunately, the tooling around doing this isn’t great, and it will almost require you to write a lot of wrapper code to make this option livable.
In walked Marten. Marten is a .NET library that uses PostgreSQL’s JSON column to extend the abilities of PostgreSQL to be both a document and a relational store. Marten solves the tooling problem I mentioned above.
Here’s how to get started with Marten.
Install the NuGet package.
dotnet add package Marten --version 4.1.0 dotnet add package Marten.AspNetCore --version 4.1.0
Now we need to configure Marten.
Now that we have Marten in place, we can start using our PostgreSQL DB as a NoSQL store.
To save a contact into the database, we could do so with the following code.
To retrieve a list of contacts.
The code to interact with Marten is very simple. They have done a great job making a simple interface that will feel very natural for C# developers.
How does Marten store data? It puts your data in a data column, and the type of data is stored in the mt_dotnet_type column.
Marten is pretty amazing. It gives us the ability to store documents in our database, but it doesn’t feel like some sort of bad hack. I recommend you give it a try.
