Recovery vs Corruption Classifier for Double-Serialized Fields and [object Object]

How it works

1. 1.Manual Trigger sends one empty item to the sample generator.
2. 2.Load Sample Records emits 4 records: once-serialized, double-serialized, containing [object Object], and an empty string field.
3. 3.Classify Serialization Depth inspects each string field and labels it corrupted, recovered with depth, or plain text.
4. 4.IF evaluates has_corrupted: output 0 corrupted, output 1 recovered/clean.
5. 5.Two Set nodes add a route field with includeOtherFields:true so classification data is preserved.
6. 6.Merge in append mode joins both routes through separate inputs.
7. 7.Build Single Report (runOnceForAllItems) returns one report item sorted by record_id, displayed by the NoOp node.

Use cases

• Audit webhook payloads whose nested objects were implicitly stringified.
• Decide which fields can be salvaged with JSON.parse and which are permanently lost.
• Measure double-serialization depth in data arriving from middleware.
• Produce a data-quality report before database insertion.

Requirements

• n8n 2.46.4 or newer; no credentials needed since the workflow is self-contained.
• For real data, replace Load Sample Records with a Webhook or HTTP Request node returning flat objects with string fields.
• Keep both Code nodes in runOnceForAllItems mode.

Setup and testing in n8n

1. Import the JSON into n8n and click Execute Workflow.
2. Check REC-001 has depth 1 and REC-002 depth 2 with the correct parsed object.
3. Check REC-003 status is corrupted and body appears in corrupted_fields.
4. Check REC-004 status is clean and the empty string field is in plain_text_fields.
5. Edge case: make the sample node return an empty array and confirm the report yields total_records = 0 without errors.

Service Value

Ready-made workflow template for automation delivery and service execution.