Handle data transformation, integration patterns, and error handling
The exam tests DataWeave fluency and disciplined error handling: typed transformations, MIME handling, error types, and correct HTTP mapping.
DataWeave rules
Declare input and output MIME types explicitly, for example output application/json with input payload application/xml. Use map, filter, reduce, groupBy, pluck, and flatten for structure; use as for coercion and default for nulls. DataSense metadata improves editor support but does not replace runtime validation. Stream large payloads and avoid converting them to strings. Handle nulls and missing keys deliberately.
Error handling
Mule errors are namespaced, such as HTTP:NOT_FOUND, VALIDATION:INVALID, and MULE:EXPRESSION. Use global and flow-level error handlers; on-error-propagate rethrows, on-error-continue consumes. A try scope isolates a subset of steps. Use raise-error to create a meaningful application error. Error handlers are evaluated in order, so put specific types before catch-alls. Map errors to HTTP status codes and return a correlation ID, not a stack trace.
Worked example
An XML order is transformed to JSON. An invalid currency code raises VALIDATION:INVALID_CURRENCY; the flow error handler maps it to HTTP 400 with a correlation ID, while unexpected errors map to 500 and are logged. Reasoning: validation failures are client errors, and preserving the correlation ID makes support possible.
Traps
A catch-all on-error-continue returning 200; swallowing errors without logging; losing the original error; mapping every error to 500; applying map to null; ignoring MIME and encoding; leaking stack traces.