This repository is derived from Anthony Nocentino's excellent examples of running SQL Server with S3 object storage integration. It is a trimmed-down and reorganized version of his work that provides a light-weight SQL Server container running alongside a minio S3 server. This allows the user to do makeshift analyses of the file formats (e.g., parquet, csv) that are supported by SQL Server's Polybase data virtualization technology.
To get started, build and run up the containers with something like:
BUILDKIT_PROGRESS=plain docker-compose up --build -d
Once the containers are running, you can upload files into the S3 service by aiming you web browser at port 9000 over an HTTPS connection on the host where the containers are running. For example, if you are running the browser on the same host where the containers are running, you can navigate to:
https://localhost:9000
The default username for the minio service is MYROOTUSER and the default password is MYROOTPASSWORD.
The ports, passwords, etc., can be modified by editing compose.yaml and Containers/SqlServerPolybase.Dockerfile.
The default password for the sa account on SQL Server is yourStrong(!)Password.
You can connect to the SQL Server instance using any of the standard management / query tools (Azure Data Studio, SQL Server Management Studio, etc.).
See Demo/demo.sql for an example query against a small CSV file that gets loaded into the S3 container at build time.