Fingerprint Dedupe and Conflict Classification Layer Around Merge Append

How it works

1. 1.Config node defines FP fields and demo samples, fanned out to two fingerprint Code nodes with type checks before trim/toLowerCase
2. 2.Merge in append mode combines both branches
3. 3.Dedupe and Classify groups by fingerprint; canonical numeric comparison separates benign duplicates from value conflicts and invalid rows
4. 4.IF routes clean records through a Set node (includeOtherFields=true) to a NoOp and suspicious/invalid to a review NoOp; the summary is fed directly from Dedupe
5. 5.Build Balance Summary emits a single item with before/after counts and a balance check

Use cases

• Purifying reports assembled from multiple tool branches
• Distinguishing benign duplicates from same-key value conflicts that may indicate double execution
• Unifying records arriving from two processing branches

Requirements

• n8n 2.44.4 or newer
• Edit the FP list only in the config node for production identity fields; for real input connect your branches directly to the fingerprint nodes
• No credentials or external calls required

Setup and testing in n8n

1. Import the workflow and run it manually from Manual Trigger
2. Inspect Dedupe: ORD-1001 benign duplicate repeatCount=2, ORD-1002 conflicting on qty and status with a metadata-only record, one invalid row
3. Verify the summary shows totalInput=7 and isBalanced=true
4. Known edge case: emptying both branchA/branchB arrays means the fingerprint nodes emit zero items, so Merge and all downstream nodes never execute and no summary is produced under executionOrder v1; the code itself never fabricates placeholder rows on empty input

Service Value

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