Parallel Array Name Inference from Parent Field with Conflict Stop-Gate

How it works

1. 1.Run the manual trigger; the first Code node emits fictional items shaped { parentField, arrays:{ key:[...] } }
2. 2.The Infer Names node slugs names, strips filler words, infers names, builds the lineage registry with naming ratios, and handles missing parents and empty arrays
3. 3.Conflicts (same name from two distinct parents) are listed and their arrays quarantined
4. 4.IF gate: zero conflicts goes through NoOp then a Set adding route=success_path with includeOtherFields; otherwise the conflict NoOp branch runs
5. 5.Merge joins branches and the final Code node emits one report item: ok or conflict_blocked with the full lineage

Use cases

• Naming arrays produced by combining many items before downstream processing
• Preventing silently misattributed columns when two paths infer the same name
• Auditing naming coverage per array for every run

Requirements

• n8n 2.44.6 with manualTrigger/code/set/if/merge/noOp/stickyNote nodes
• No credentials or network calls; inference rules live inside the Code node
• For real use, replace the sample emitter with input of the same schema: { parentField: string, arrays: object }

Setup and testing in n8n

1. Import and run Manual Trigger; expect status=conflict_blocked because customer_data and customer_details both infer the name customer
2. Inspect quarantinedArrays, cleanNamedArrays and lineage in Build Final Report output
3. Remove the two conflicting sample items and rerun to get status=ok with naming ratio 1
4. Edge cases: empty arrays and an item without parentField appear in skipped with lineage status skipped_*

Service Value

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