n8n Payload Mapping Repair: Explicit Results for Every Input

AI-assisted, self-built synthetic portfolio demonstration. All records are invented. No client, live webhook, external API, credential or production system was involved.

A three-node n8n sample supplied an email inside body.email. Its Code node instead read a nonexistent top-level email and called .trim() on it. The batch stopped with TypeError: Cannot read properties of undefined (reading 'trim') before producing the expected results.

This is a payload-mapping and validation example. It is separate from my API-retry and pagination demonstrations.

The repaired Code node follows an explicit input contract:

Read the email from item.json.body.email.

Require an object body and a string email. Reject JSON-string bodies rather than guessing their meaning.

Trim surrounding whitespace and lowercase under this demo's agreed normalization rule.

Return one accepted or rejected record per input, with an explicit reason for rejection.

Preserve the original input index through n8n's pairedItem field.

With the two-record sample, DEMO@EXAMPLE.COM becomes an accepted demo@example.com. The null email becomes a rejection with the reason email must be a string. A malformed record no longer stops the valid record in that sample batch.

The test harness executes the exact JavaScript extracted from the exported Code node, rather than a separate implementation. The captured November 5, 2026 run passed 10 of 10 Node.js VM test groups, with zero failed groups:

Normalize the valid sample and reject its null email.

Handle empty input.

Reject a missing body.

Reject null and array bodies.

Reject a string body without silently parsing it.

Reject missing and numeric emails.

Reject blank and whitespace-only emails.

Reject the tested malformed address shapes.

Normalize the tested plus-tag and subdomain example.

Retain duplicate inputs as separate records.

The harness also checks that the inputs are unchanged, output records preserve input-index linkage, the exports parse, both workflows are inactive, and no node has credentials.

There is separate, earlier engine evidence: on November 3, 2026, both workflows were imported and executed through the official n8n 1.121.3 CLI. The broken workflow reproduced the error. The repaired Normalize Email node reported success and returned the two expected records. That engine check covers the two-record sample only; the 10 edge-case groups ran in Node.js. The saved CLI result has finished: true and node-level success, while its aggregate status field still reads running.

This sample uses a Manual Trigger, not a live webhook. The address check tests basic shape, not full RFC compliance or deliverability. Lowercasing is an explicit demo rule, not a universal recommendation for every email system. Duplicates are retained. No retry handling, cross-request idempotency, persistence, authentication or downstream action was tested. Consumers must branch on the returned status before taking action.

The newer n8n 2.46.4 package could not be installed in the original environment because of a platform dependency mismatch. No compatibility claim is made for that version. This case study makes no claim about paid client work, production reliability, time saved or revenue impact.

For a real workflow repair, first agree the payload shape and rejection policy. Reproduce the failure with minimized fixtures, test the exact exported code, then verify the relevant execution path and downstream behavior in a safe n8n environment.

Service: Bounded n8n workflow repair