Salesforce MuleSoft Developer: Transforming Data with DataWeave

Script anatomy

A DataWeave file starts with %dw 2.0, optionally import/include statements, then declarations (var, fun, type) separated from the output expression by ---. Use fun for reusable logic and type for custom schemas. The expression sees payload, vars, attributes, and p('property').

Selectors and operators

Select with .key, ."key with space", [index], and .*key for repeated keys. ++ concatenates strings, arrays, and objects; -- removes keys or values; default supplies a fallback when the left side is null; as coerces to a type.

Core functions, by output shape

• Same number of elements: map (element $, index $$), mapObject (value $$, key $), filterObject.
• Fewer elements: filter.
• One value: reduce.
• One per group: groupBy then mapObject or pluck.
• Objects to arrays: pluck, entriesOf; arrays to objects: zip, flatten, distinctBy, orderBy, sizeOf.

Traps

• In mapObject, $$ is the key and $ the value - the reverse of common intuition.
• map cannot iterate an object; use mapObject or pluck.
• groupBy returns an object of arrays with no guaranteed ordering.
• Coercion failures raise errors; guard with default or pattern matching (case / if).

Worked example

payload groupBy $.customerId pluck ((orders, id) -> { id: id, count: sizeOf(orders), total: orders reduce ((o, acc=0) -> acc + o.amount) }) returns one summary object per customer, with the group array as the first lambda argument and the key second.

Rule of thumb

Choose the function from the output shape you need, not from habit: many-in/many-out is map, many-in/fewer-out is filter, many-in/one-out is reduce.