HydraData is an embeddable, script-driven data pump for .NET 10. It discovers ordered C# step scripts, validates them before execution, and runs them against SQL Server or PostgreSQL with scoped state, transactions, controlled IO, and host-friendly exit codes.
HydraData is intended to be suitable for public GitHub use as an OSS project under the MIT license.
The root README is intentionally short. Detailed setup, scripting, operations, and embedding guidance belongs under docs/.
- .NET 10 SDK
- Git
- SQL Server or PostgreSQL for real database steps
The hello trial below does not open a database connection, but the host still
needs a valid connections.xml with at least one connection entry. When scripts
do not select a connection, HydraData uses the first declared entry; naming it
default is a recommended convention.
dotnet restore hydradata.slnx
dotnet build hydradata.slnx
dotnet test hydradata.slnx --filter 'FullyQualifiedName!~IntegrationTests'HydraData.Host reads appsettings.json, connections.xml, and script folders from the current working directory. From the repository root, create scripts/, input/, and output/, then add these files.
scripts/01_10_hello.cs
// @name: Hello HydraData
// @description: Minimal smoke test
// @haltOnError: true
// @unsafe: false
Note("HydraData executed the hello step.");
Print("Hello from HydraData.");
return Ok("Hello step completed.");appsettings.json
{
"Pump": {
"WorkspaceBase": "./_runs",
"AllowUnsafeDirectAccess": false,
"ReadAllowlist": [ "./input" ],
"WriteAllowlist": [ "./output" ],
"StepTimeoutSeconds": 120,
"RunDirRetentionDays": 14,
"ScriptFolders": [ "./scripts" ],
"ConnectionsFile": "./connections.xml"
}
}connections.xml
<ConnectionStrings>
<ConnectionString targetSystem="MSSQL" name="default">
<Parameters>
<Parameter key="Server" value="localhost" type="String" />
<Parameter key="Database" value="stage" type="String" />
</Parameters>
</ConnectionString>
</ConnectionStrings>Run the host:
dotnet run --project src/HydraData.Host/HydraData.Host.csprojA successful run returns exit code 0 and writes run artifacts under _runs/. Edit connections.xml before using database APIs such as Query, Scalar, Execute, or BulkInsert.
Use HydraData.Engine when you want to embed the pump in another .NET process. The main entry point is IPumpEngine, implemented by PumpEngine.
Typical embedded flow:
var options = new PumpOptions(workspaceBase, folders);
var engine = new PumpEngine(options);
var scripts = new DiscoveryService().Discover(["./scripts"]);
var registry = ConnectionRegistry.Load("./connections.xml");
var connections = new ConnectionDirectory(registry);
var externCtx = ExternContext.FromValues(new Dictionary<string, object?>());
var validation = engine.Validate(scripts, externCtx, connections);
if (!validation.IsValid)
return 1;
var report = await engine.ExecuteAsync(scripts, externCtx, connections, ct: cancellationToken);
return report.ExitCode;See the public docs linked above for the full API surface, script conventions, and operational guidance.
HydraData is licensed under the MIT license. See LICENSE.txt.