From WMS to UFIDA U8: A Practical Breakdown of the Other-Inbound Sync Strategy
What This Strategy Solves
In retail-and-manufacturing operations, warehouses record non-purchase, non-sale inbound events—stocktake surpluses, transfers, assembly/disassembly, and gift receipts—into the "Other-Inbound" document. These documents must flow into the ERP to form the book-side inventory value. We use Qeasy (the Qeasy Data Integration Platform) to move these documents incrementally from the source system to the target system, with business types and codes normalized in an intermediate layer.
Data Flow and Field Mapping
The overall flow is "Source system → Qeasy intermediate layer → Target system". Key field mappings are as follows:
| Meaning | Source Field | Intermediate Standardization | Target Field |
|---|---|---|---|
| Document number | Original document number | Unified prefix + source identifier | Other-Inbound number |
| Business type | Inbound type enum | Business type code table | Document type |
| Warehouse | Warehouse code | Warehouse mapping table | Warehouse code |
| Item code | SKU code | Unified master-data code | Item code |
| Quantity | Quantity field | Value normalization (sign) | Quantity field |
| Price/Amount | Amount field | Tax-inclusive/exclusive split | Unit price, Amount |
| Remarks | Free text | Truncation and masking | Remarks |
| Creator/Time | Operator, timestamp | Time zone unified to UTC+8 | Creator, Audit time |
Centralized code mapping is the most common pattern among Qeasy customers: keep warehouses, items and business types in a global mapping table, so adding a new store or adjusting a business type later only requires editing one place.
How to Configure on Qeasy
- Data source access: Use the platform's standard adapter or SDK on the source side; use the financial adapter pointing to the target ledger on the target side.
- Data models: Build one data model each for the source and target tables, with fields corresponding one-to-one in the intermediate layer.
- Transformer: Use a script transformer for the header to handle dates, code mapping and document-number concatenation; loop through the body to process amount and quantity per inventory row.
- Write-back: Write the intermediate data back to the target system's "Other-Inbound" interface, paying attention to document status fields and mandatory validations.
- Exception handling: Enable the platform's built-in retry and configure a failure queue; common cases like "item code does not exist" or "warehouse disabled" go through a manual document-repair channel.
On customer sites we usually split the header and body into two phases: first get the header running, confirm document number, warehouse and business type are correct, then release the body loop. This staged approach keeps the troubleshooting chain short when issues arise.
Implementation Steps
- Set the incremental cursor: Use the source system's last-modified time as the cursor (typically the inbound document's modification time); first backfill the last 3 months as a baseline.
- Full trigger: Use Qeasy's full schedule to fire an initial run once, recording each document's source ID and target receipt number to build a reconciliation table.
- Schedule frequency: Poll increments every 15 minutes by default; if near-real-time is required, shorten to every 5 minutes when the source can open change push, combined with a trigger-style callback.
- Reconciliation and replay: Run a "source vs target" reconciliation script once every early morning; differences enter a repair queue, replayed automatically by the platform or manually.
Pitfalls Review
- Business type enums not unified. Source and target name "surplus / transfer / assembly-disassembly" differently, with no code table in the intermediate layer, causing mismatched names after launch. The safe approach: maintain all business types only in the global mapping table.
- Amount sign errors. Other-Inbound documents usually increase inventory, but mixing in red-letter documents without sign normalization causes the target system to also write returns as positive. Always judge the sign of quantity/amount by business type inside the body loop.
- Missing codes block writes. If an item or warehouse is not enabled in the target system, the write will hang the whole document. A typical pitfall is relying on the other side to maintain master data manually. Qeasy's typical response: place a "code pre-validation" before the transformer, filter out non-enabled items before writing the main document.
- Time zone and document time drift. Source and target systems have inconsistent time zones, causing reconciliation to drift by hours. Force UTC+8 unification in the intermediate layer.
- Mixing full and incremental overwrites data. Re-running full triggers the target system's overwrite logic. The safe approach: use full only for the first initialization, all subsequent runs use incremental; if a full re-run is required, clear the target side first before triggering.
Applicable and Non-applicable Scenarios
Applicable: Non-purchase, non-sale inbound scenarios such as stocktake surpluses, transfers, assembly-disassembly and gift receipts that need to form ERP book inventory. Not applicable: Use cases requiring second-level real-time when the source can directly open change push (use a real-time event bus instead); nor suitable for scenarios requiring complex approval flows and multi-level signatures before booking.