Practical Tutorial: Syncing Guanyi Sales Delivery Orders to Kingdee Other-Outbound Orders (Online Repair Scenario)
What This Strategy Solves
For a retail business running its online repair workflow in Guanyi Cloud, every completed repair order generates a sales delivery order in Guanyi, marking the physical outbound of the repaired or replaced item. On the finance side, Kingdee Cosmic Star needs each of those events recorded as an other-outbound order, so that cost accounting and inventory reduction stay accurate. When the two sides drift apart, you end up with inventory mismatches and repair costs that never land in the right accounting period.
The strategy itself is straightforward: turn Guanyi's sales delivery orders into other-outbound orders in Kingdee, one in and one out, so the books balance on both sides.
Data Flow and Field Mapping
The end-to-end path is: Guanyi Cloud (source) → Qeasy Data Integration Platform (middleware) → Kingdee Cosmic Star (target).
The middleware handles encoding mapping, field conversion, and document state handoff. The key fields (focused on the ones that cause the most engineering pain) are listed below.
| Business Meaning | Guanyi Sales Delivery (Source) | Kingdee Other-Outbound (Target) | Notes |
|---|---|---|---|
| Document number | Delivery order no. | Document no. (FBillNo) | Idempotent write in middleware |
| Document date | Delivery time | Business date | Timezone follows the customer |
| Customer code | Customer code | Customer (FCustomerId) | Encoding mapping centralized |
| Warehouse code | Outbound warehouse | Warehouse (FStockId) | Warehouse master must sync first |
| Item code | SKU code | Material (FMaterialId) | Material master must sync first |
| Quantity | Quantity | Actual qty (FQty) | Units must be unified |
| Price / amount | Sales price | Unit price / amount | Often 0 in repair scenarios |
| Document type | / | Business type (other-outbound) | Hard-coded on target |
Tip: keep all encoding mappings centralized in Qeasy, never scattered across per-strategy scripts. We have seen customers put mappings into each strategy's transform script and, three months later, the maintenance team simply cannot find or change them.
How to Configure on Qeasy
Landing this strategy on Qeasy Data Integration Platform comes down to four building blocks: source read, field mapping, target write, and scheduling with alerting.
- Source connector: Configure the Guanyi sales delivery order interface and filter on the "approved" status, so draft documents are never pushed downstream.
- Target connector: Configure the Kingdee other-outbound save interface. The business type is fixed to "other-outbound", and dimensions like organization, inventory organization, and warehouse are passed as parameters.
- Field mapping: Do the alignment in Qeasy's mapping canvas. When units disagree, convert on the source side rather than relying on the target to fix it up. If the target conversion fails, the document gets stuck inside Kingdee.
- Exception handling: The standard Qeasy pattern is to push failed writes into a "retry queue" for the on-duty engineer to handle, instead of bouncing them back to the source system.
Implementation Steps
In a real engagement, we usually move through three phases: incremental start point, full backfill, and scheduling cadence.
Phase 1: Confirm the incremental start point. The very first thing on-site is not writing scripts, it is deciding the incremental start point. Online repair volumes are modest, so using "approved orders in the last 7 days" as the start point is usually the safest call. Anything older than that, if the volume is small, is handled by manual entry; if the volume is large, we kick off a separate historical backfill.
Phase 2: Full backfill. For the stock of data after the start point, manually trigger a full sync in Qeasy. The moment the backfill finishes, reconcile: Guanyi delivery count vs Kingdee other-outbound count, and the variance must be zero. If it is not, investigate two typical causes first — duplicate documents and unmapped codes — before touching any script.
Phase 3: Scheduling cadence. Online repair needs decent real-time behavior but is not as critical as standard sales orders. Our common configuration is: pull every 10 minutes, then batch-write to the target. Customers with different store footprints can tune it to 5 or 15 minutes. Going to 1 minute will saturate Guanyi's query interface and gain nothing.
Pitfalls From Real Projects
Here are the landmines we have actually hit on customer sites.
- Scattered encoding mappings. The most typical mistake is keeping the warehouse / customer / material code tables inside each strategy's own script. Three months later, an org change means digging through every strategy one by one. The safe pattern is to maintain one shared encoding-mapping table in Qeasy and have every strategy reference it.
- Header and body submitted together. Kingdee's other-outbound order has a header-plus-lines structure. If Qeasy only writes the header, or the body is empty, or the line count is wrong, the target rejects it. We have customers adopt a "header first, body second" staging pattern: write the header, then the body, and on failure do not leave a half-baked document behind.
- Status filter missed. Pushing draft, pending, or unapproved deliveries downstream produces a flood of "other-outbound" drafts in Kingdee, and finance month-end close becomes a mess. Filtering status on the source side is basic hygiene, but it is easy to forget.
- Base unit vs transaction unit confusion. A Guanyi SKU might be in "piece" while the Kingdee material is in "PCS". They look the same until write time, when unit conversion fails. The safe approach is to convert to the target's base unit on the source side before pushing.
- Timezone drift. For cross-region customers, the source timestamp carries one timezone and the target stamps another, and document dates can shift by a day. Pinning a single timezone (usually the customer's HQ location) in the configuration is the easiest way out.
When This Applies and When It Does Not
Applies to: online repair, after-sales replacement, sample outbound, and any other "non-sales but inventory-out" business where daily volume is in the hundreds to low thousands and finance needs a full audit trail.
Does not apply to: standard sales outbound (use sales-outbound order instead), cross-org transfers (use transfer order), and scenarios where the source documents are unstable, frequently voided, or reversed. In the last case, reconciliation logic gets messy and the right first step is to clean up the source side before automating the integration.