Stream represents ordered set of events. Simply speaking stream is an log of the events that happened for the specific aggregate/entity.
As truth is in the log, so in the events - stream can be also interpreted as the grouping, where key is stream id.
Id - needs to be unique, so normally is represented by Guid type.
It's also common to add two other columns:
Type- information about the stream type. It' mostly used to make debugging easier or some optimizations,Version- auto-incremented value used mostly for optimistic concurrency check.
See more in Greg Young's "Building an Event Storage" article
In this exercise you should write the code that will create proper Streams table based on the requirements described in the Overview section. You need to go to EventStore class - it has already some boilerplate code and structure defined. You need to place your logic of creating table into CreateStreamTable method.
To verify if your code you need to run tests from Exercise01CreateStreamsTable file. It contains precreated tests that will turn green when code is written properly.
It's suggested to use Dapper to perform the database calls. Dapper is a simple object mapper for .Net provided by and used in StackOverlfow. It simplifies the mapping between the raw SQL commands/queries and the objects.
To execute command without the result you can use Execute method. It will automatically open connection, so you don't need to do it manually.
As the database we'll use Postgres. It's an open source, mature and well designed and performant database.
It also has great support for JSON (probably the best from the popular relational databases) - see more here or there.
For the syntax of the table creation you can refer to the official documentation.
You can find mappings of the Postgres column types to the .NET types in Npgsql documentation
Note: If you prefer to use different database feel free to do that, all of the exercises and samples of creating own Event Store should be easily transferable to other database engines.