Classification Index Tally Reconciliation Across Merge Branches

How it works

1. 1.Run the manual trigger to emit fictional records with two sub-arrays plus one invalid record
2. 2.Branch A cross-product explodes arrays, promoting each key to a column, and drops one row to simulate silent loss
3. 3.Branch B counts, per category value, how many source records contain it as the reference tally
4. 4.An append merge collects both branches, each on its own input
5. 5.A code node builds the inverted index and balances each value count against the reference
6. 6.IF routes drift=0 items to the report and deviated items through the NoOp drift checkpoint
7. 7.Report branches are joined and one final item is emitted: sorted index, reconciliation, drift report, overall status

Use cases

• Auditing silent row loss when exploding arrays across multiple merged branches
• Building a trustworthy sortable classification index before committing results
• Screening invalid records before decomposition

Requirements

• n8n 2.44.6 with code, merge, if, noOp, stickyNote and manualTrigger nodes
• No credentials needed; fictional test payload is embedded
• To use real data, replace the sample Code node with a real source using schema {record_id, <categoryKey>: string[]}

Setup and testing in n8n

1. Import the JSON into n8n and execute 'Run Demo'
2. Verify the final report shows channel/phone with drift -1 (R-003 row lost)
3. Test the healthy path by removing the simulation line in Branch A so status becomes balanced
4. Test edge cases with empty records or missing record_id; they appear in invalid_records

Service Value

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