It's often struct me how fundemental data structures such as stacks and queues end up being implemented in an ad hoc fashion as part of relational data modeling, and I've wondered what it would look like to implement various data structures generically in a relational data store. Meanwhile, I am starting a new job soon which uses C# and Azure after spending the past 6 years mostly using F# and AWS. Not a huge leap, but during my gap time between jobs (4 weeks!), I've decided to take on a project to sharpen my C# skills and get my hands dirty with Azure, and this seemed like a good project to structure my efforts around.
We assume Linux OS with .NET 6, Docker, and SQL Server 2019 for Linux installed locally
The following environment variables must be defined: DSTRUCT_DB_SERVER, DSTRUCT_DB_USER, DSTRUCT_DB_PASSWORD
- Run
./sql/run.sh create-db.sqlto drop and recreate theDStructdatabase plus tables and stored proceedures. - (optional) Run
./sql/run.sh create-test-data.sqlto create some test data. - (bonus) Run
./sql/run.shwithout arguments to start an interactivesqlcmdsession.
There is also a Makefile in the sql folder which adds some convenience around these commands.
Run docker compose up to start the www and api projects running on ports 8080 and 9090 respectively.
You could call this a mono-repo, with three distinct projects organized under subdirectories:
wwwis a C# / .NET 6 / ASP.NET MVC project that provides a homefront for the project with a blog documenting my adventuresapiis a C# / .NET 6 / ASP.NET WebAPI project that provides a RESTful interface for our persistent data structuressqlis a T-SQL script folder that implements our data structures
We plan to use Docker and AKS to deploy the two apps to www.dstruct.dev and api.dstruct.dev domains, and Azure SQL as our DB server.