Sales Outbound Order Sync in Practice: A Supply Chain Implementation from Jushuitan to Kingdee
What This Strategy Solves
In a real engagement with a retail company, Jushuitan acts as the order aggregation point for e-commerce and store channels, while Kingdee Cloud Xingchen serves as the back-end finance and supply chain system of record. Once a sales outbound order is approved for shipment in Jushuitan, it must flow into Kingdee promptly—otherwise financial closing, cost calculation, and receivables confirmation all get delayed. This strategy addresses the "ship-now, book-now" requirement: pushing sales outbound orders from Jushuitan to Kingdee at the document level, keeping inventory and revenue consistent across both sides.
Data Flow and Field Mapping
The flow is straightforward: Jushuitan (source) → Qeasy integration platform (middleware for cleansing, mapping, transformation) → Kingdee Cloud Xingchen (target). The middleware does not own business entities; it only routes and translates.
Key field mapping (typical scenario; actual fields depend on each system's metadata):
| Business Meaning | Jushuitan (Source) | Kingdee (Target) | Handling Notes |
|---|---|---|---|
| Document No. | io_id / io_no | FBillNo | Generated on Kingdee side; raw Jushuitan number stored in remarks |
| Document Date | io_date | FDate | Normalize timezone and date format |
| Customer Code | shop_id / buyer | FCustomerID | Maintained centrally in Qeasy "code mapping" |
| Material Code | sku_id | FMaterialID | Shares the same mapping table as the material sync strategy |
| Warehouse | warehouse_id | FStockID | Warehouse master must be synced before outbound orders |
| Quantity, Price | qty, price | FQty, FPrice | Numeric precision unified to two decimals |
| Header Remark | remark | FNote | Truncate to target system's allowed length |
How to Configure on Qeasy
On the Qeasy platform, this strategy is modeled as a "document flow," not a "data table flow." There are three configuration points:
First, on the source side, use the Qeasy Jushuitan connector and subscribe to increments for the "sales outbound order" document type. Use Jushuitan's internal document number as the primary key, decoupling it from external order numbers. Second, in the middleware, use Qeasy's transformation components for field mapping. The three critical master data entities—customer, warehouse, and material—are handled with a "centralized code mapping" pattern: a single mapping table that other material or customer strategies can reuse. Third, on the target side, use Kingdee's save interface that supports header and line items together, and write line items split out one by one.
One point that's easy to miss: Kingdee Cloud Xingchen validates header fields more strictly than line items. If customer, currency, or date types are wrong, the entire document is rejected. We recommend adding a "pre-validation" step in the Qeasy transformation so that required header fields are validated before writing to the target.
Implementation Steps
We usually split go-live into three phases:
Phase 1: Align the incremental starting point. On Qeasy, position the source cursor at "outbound orders approved for shipment since 00:00 on the go-live date," and do a small backfill first (a common approach is to take the last 7 days of data) to validate mapping and the target interface. This phase does not depend on historical full loads; it only establishes a starting point for subsequent increments.
Phase 2: Full trigger and reconciliation. When the historical outbound order volume is manageable (e.g., orders within the last 3 months), trigger a one-time full backfill from Jushuitan using a time window. Once the full load is complete, we switch the task on Qeasy to an "incremental + full dual-track" mode: full loads handle catch-up, increments handle daily sync, with deduplication by document number.
Phase 3: Steady-state scheduling. For outbound orders, we recommend scheduling every 15 minutes, balancing real-time needs and interface pressure. Qeasy supports schedule-expression triggers, with failed tasks entering a retry queue—typically 3 retries is enough to cover occasional network jitter. If business requires higher real-time, this can be shortened to 5 minutes, but the write concurrency on the Kingdee side should be evaluated first.
Pitfalls Review
Pitfall 1: Customer master not in place yet. The outbound order write fails with "customer does not exist." The root cause is that the upstream customer sync strategy and the outbound order strategy have no dependency. The safe approach is to set a soft dependency on Qeasy from the outbound order strategy to the customer sync strategy—when target validation fails, the task enters a to-do state rather than throwing an error directly.
Pitfall 2: Warehouse IDs don't match across systems. Jushuitan uses text warehouse codes, while Kingdee uses numeric IDs; passing them directly triggers a type error. We adopt a "centralized code mapping" approach—all warehouses are maintained in one mapping table on Qeasy. New warehouses only require updating the mapping table, not the strategy.
Pitfall 3: Line items get truncated. During a promotion, one outbound order had over 80 SKUs, but Kingdee's interface only accepts 30 lines by default, so the rest were lost. We recommend adding line splitting in the Qeasy transformation step based on the target interface limit, with automatic document splitting when the threshold is exceeded.
Pitfall 4: Timezone and date boundaries. During cross-day scheduling, if the cursor uses "up to 23:59:59 yesterday," orders from the last 5 minutes of the day are missed. The safe approach is to add a small overlap (e.g., 5 minutes) and then deduplicate using the target system's document date.
Pitfall 5: Idempotency not guaranteed. Retries cause duplicate documents. A typical mistake is not using the document number as a deduplication key. In the Qeasy transformation step, we recommend writing the Jushuitan document number into a Kingdee remark field so duplicates are obvious during manual reconciliation.
Suitable and Unsuitable Scenarios
This strategy fits e-commerce retail, multi-store distribution, and scenarios where front-end and back-end systems are heterogeneous and need sales outbound synchronization. It is not suitable for pure internal ERP-to-ERP reconciliation, cross-entity consolidated reporting, or scenarios requiring granular batch-level cost accounting—the latter typically requires a dedicated cost calculation strategy rather than the outbound order itself.