Parse JSON in Make: objects, arrays, missing values and output bundles

Choose one explicit input shape

The example is an object containing a request identifier and an array of items: {"request_id":"REQ-401","items":[{"sku":"A01","qty":2}]}. Some webhook or HTTP outputs are already structured and can be mapped directly. Do not add another parser merely because the data resembles JSON. Identify whether the source value is text, a collection or an array first.

Distinguish JSON text from structured fields

Parse JSON is for text encoded as JSON. If the upstream output is already a collection, inspect its fields before converting it back to text. Broken quotation marks, escapes or literal newlines can make text invalid. Adding braces by eye is unreliable. Keep the original input available for diagnosis while avoiding unnecessary copying of sensitive payloads into broad logs.

Define and inspect the data structure

Generate a Data structure from the sample and confirm request_id is text, items is an array and qty is numeric. A sample does not automatically define every optional field your real input might contain. State the required fields explicitly, inspect a manual output and then connect downstream modules. Check both the mapped value and its type, especially where spreadsheet or API input may represent numbers as text.

Separate root arrays from nested arrays

With an object root, items remains a nested array that you can pass to an Iterator when each item needs processing. A root array parsed by Parse JSON can produce separate output bundles for its elements. Inspect the output before adding an iterator for the same purpose. Define the expected number of records and notifications so an accidental extra expansion is detectable.

Handle missing fields, null and empty strings deliberately

A missing memo, memo set to null and memo set to an empty string can imply different update actions. Reject a missing request identifier while preserving an existing optional memo when no update was supplied. Decide whether a text quantity such as "2" is accepted or rejected, and whether zero or negative quantities are valid. Successful parsing does not validate those business rules.

Build outgoing JSON from structured fields

Use Create JSON or the HTTP data-structure body option to map outgoing values. Concatenating raw strings into JSON breaks when a subject contains quotes or newlines. Include those characters in a proposed test input and compare the receiving field with the original text. Keep authentication in the connection rather than inserting it into the business payload or a shared sample.

Reconcile outputs and error cases

Try two items, an empty items array, a missing identifier and syntactically invalid JSON. With an empty array, check whether zero downstream bundles also prevent an intake-level record that you still need. Validate identifiers and quantities separately from syntax. Count expanded writes and notifications when estimating usage. This guide defines checks to run; it does not claim measured execution or a verified importable blueprint.

• Connect a website backend to a Make custom webhook
• Iterator and Aggregator in Make: arrays, bundles and summary outputs
• Handle Make errors: execution history, incomplete runs and safe retries