README.md

AzureFunctions.TestFramework.Sql

Azure SQL Trigger, Input, and Output binding support for the Azure Functions Test Framework.

Installation

dotnet add package AzureFunctions.TestFramework.Sql

Supported bindings

Binding	Attribute	Description
[SqlTrigger]	Trigger	Receives a batch of row changes from SQL change tracking
[SqlInput]	Input	Reads rows from a SQL table or view via a query
[SqlOutput]	Output	Upserts rows into a SQL table — captured via FunctionInvocationResult

SQL Trigger

Use InvokeSqlAsync to simulate a SQL change-tracking trigger with a batch of row changes.

Strongly-typed changes

using Microsoft.Azure.Functions.Worker.Extensions.Sql;

var changes = new[]
{
    new SqlChange<Product>(SqlChangeOperation.Insert, new Product { Id = 1, Name = "Widget" }),
    new SqlChange<Product>(SqlChangeOperation.Update, new Product { Id = 2, Name = "Gadget" })
};

var result = await host.InvokeSqlAsync("ProcessSqlChanges", changes);

Assert.True(result.Success);

Raw JSON

var json = """[{"operation":0,"item":{"id":1,"name":"Widget"}}]""";

var result = await host.InvokeSqlAsync("ProcessSqlChanges", json);

Assert.True(result.Success);

SQL Input Binding

Register fake rows via the builder so they are injected automatically for every invocation:

var host = await new FunctionsTestHostBuilder()
    .WithFunctionsAssembly(typeof(MyFunction).Assembly)
    .WithHostBuilderFactory(Program.CreateHostBuilder)
    // Inject rows for [SqlInput(commandText: "SELECT * FROM Products", commandType: CommandType.Text, ...)]
    .WithSqlInputRows("SELECT * FROM Products",
        new List<Product>
        {
            new Product { Id = 1, Name = "Widget" },
            new Product { Id = 2, Name = "Gadget" }
        })
    .BuildAndStartAsync();

Use WithSqlInputRows<T>(commandText, T row) to inject a single row, WithSqlInputRows<T>(commandText, IReadOnlyList<T> rows) to inject a list, and WithSqlInputJson(commandText, json) to inject raw JSON.

The commandText must exactly match the value declared in the [SqlInput] attribute.

Function example

[Function("ReadSqlProducts")]
public void Run(
    [QueueTrigger("products-queue")] string queueMessage,
    [SqlInput(commandText: "SELECT * FROM Products",
              commandType: System.Data.CommandType.Text,
              connectionStringSetting: "SqlConnection")] IEnumerable<Product> products)
{
    foreach (var product in products)
        _logger.LogInformation("Product: {Name}", product.Name);
}

SQL Output Binding

Output bindings are captured automatically via FunctionInvocationResult:

var changes = new[] { new SqlChange<Product>(SqlChangeOperation.Insert, new Product { Id = 1, Name = "Widget" }) };
var result = await host.InvokeSqlAsync("UpsertSqlProduct", changes);

Assert.True(result.Success);
var written = result.ReadReturnValueAs<Product>();
Assert.Equal(1, written?.Id);

Function example

[Function("UpsertSqlProduct")]
[SqlOutput(tableName: "Products", connectionStringSetting: "SqlConnection")]
public Product? Run(
    [SqlTrigger(tableName: "Changes", connectionStringSetting: "SqlConnection")]
    IReadOnlyList<SqlChange<Product>> changes)
{
    return changes.FirstOrDefault()?.Item;
}

Testing across all four flavours

Add the SQL package reference to your test project and all four function-app test flavours:

<PackageReference Include="AzureFunctions.TestFramework.Sql" />

See the 4-flavour matrix test pattern for the concrete test class structure.