WJb Docs - Intermediate

📨 Passing Data

Data moves between actions through the payload of the next job.

There is no hidden context or ambient state.

---

Basic Pattern

The current action creates a new payload and passes it to the next action:

public override async Task<IActionResult> ExecuteAsync(
    OrderInput input, CancellationToken ct)
{
    // process the order...
return await NextAsync<SendEmailAction>(
        new EmailInput
        {
            To = input.CustomerEmail,
            Subject = "Order Confirmation",
            Body = $"Your order {input.OrderId} has been received."
        });
}

The EmailInput object becomes the payload of the new job.

---

Strongly Typed Inputs

Each action declares the data it expects:

public sealed class EmailInput
{
    public string To { get; init; } = string.Empty;
    public string Subject { get; init; } = string.Empty;
    public string Body { get; init; } = string.Empty;
}

WJb automatically deserializes the job payload into this type.

---

Passing Results Forward

You can enrich or transform data before passing it on:

public override async Task<IActionResult> ExecuteAsync(
    ImportInput input, CancellationToken ct)
{
    var customers = await ImportCustomersAsync(input, ct);
return await NextAsync<ProcessCustomersAction>(
        new ProcessInput
        {
            Customers = customers,
            ImportedAt = DateTime.UtcNow
        });
}

---

Multiple Next with Different Data

Each command carries its own payload:

return Results.Next(
    JobCommands.Next<LogAction>(
        new LogInput { Message = "Import finished" }),
    JobCommands.Next<NotifyAction>(
        new NotifyInput
        {
            Email = input.AdminEmail,
            Count = customers.Count
        }));

---

What Is Not Shared Automatically

• Local variables
• Private fields of the action
• Ambient context or HttpContext
• Previous job results (unless you explicitly pass them)

Everything that the next action needs must be placed into its payload.

---

Best Practices

• Keep payloads small and focused
• Prefer strongly typed models over anonymous objects for public workflows
• Pass only the data the next action actually needs
• Avoid putting large files or binary data directly into payloads (use storage references instead)

---

Key Point

Data flow is explicit.

The current action decides exactly what information the next action will receive.