Practical Tutorial: Syncing Kingdee Distributed Transfer-In Documents to Wangdiantong Purchase Inbound

What This Strategy Solves

In the supply chain of a retail enterprise, Kingdee Cloud Cosmos acts as the back-end ERP, generating distributed transfer-in documents and recording accounting entries. Wangdiantong, as the front-end warehouse and purchasing execution system, requires inbound vouchers to proceed with put-away and reconciliation. Since the two systems use different document conventions and transfer-in documents cannot directly serve as purchase inbound, a strategy is needed to reliably convert distributed transfer-in documents into purchase inbound documents, sitting downstream of purchase orders and purchase receipt notices.

Data Flow and Field Mapping

The overall direction is B_TO_A. Kingdee Cloud Cosmos (source system B) exports distributed transfer-in documents. After conversion and code mapping in the Qeasy data integration platform, the data is written into Wangdiantong's (target system A) purchase inbound interface.

Key field mapping (typical items, actual fields depend on both interfaces):

• Document number: source document number → target document number (both sides require uniqueness; usually source number with a prefix is written).
• Warehouse code: source ERP warehouse → Wangdiantong warehouse code (maintained in a mapping table).
• Vendor code: source ERP vendor primary key → Wangdiantong vendor code.
• Item code: material code → SKU code (cross-system item master data must be governed in advance).
• Quantity, unit price, amount: carried through directly without conversion.
• Business date: source document inbound date → inbound business date.
• Remarks/notes: extension fields are transparently transmitted.

For code mapping, we usually adopt a "centrally managed + version-traceable" approach, maintained uniformly in Qeasy. When material or vendor codes change later, only one place needs to be updated.

How to Configure in Qeasy

Typical configuration points for this strategy in the Qeasy platform:

1. Data source registration: Connect Kingdee Cloud Cosmos as the source, using its distributed transfer-in document query interface; connect Wangdiantong as the target, locating the purchase inbound document write interface.
2. Extraction strategy: Incremental pull, filtered by document modification time plus document status, ensuring only approved documents are taken, avoiding draft contamination downstream.
3. Transformer: Complete field mapping, unit validation, and code lookup in Qeasy's transformer. For headers and line items, adopt a "header first, lines phased" stable writing approach—write header info first, then batch-submit line items—to avoid partial failures breaking the entire document.
4. Target write: Wangdiantong's write call requires mandatory fields like warehouse, vendor, and SKU. Any unmapped codes will be intercepted by Qeasy and logged, preventing dirty data from flowing downstream.
5. Exception and retry: Configure retry and alerts in Qeasy. Common errors such as missing codes or empty mandatory fields will go directly into the exception queue for manual investigation.

Implementation Steps

We recommend a three-step approach, avoiding jumping straight into full sync which can confuse both sides:

• Step 1: Align incremental start point. Freeze a starting timestamp first, and sync only new or changed distributed transfer-in documents after that point. Run for 3–5 days, then verify that Wangdiantong's inbound document count and amount match the source system.
• Step 2: Full sync trigger. Once incremental is stable, trigger a full backfill via Qeasy to fill in documents needed before the start point. The full sync phase should be scheduled during off-peak hours, with batch submission to reduce pressure on the target system.
• Step 3: Scheduling frequency. After stabilization, bring the strategy into scheduled execution. A common practice is to run incremental every 15–30 minutes, balancing real-time performance without obvious pressure on the source system. Full sync is triggered monthly or on demand.

Lessons Learned

• Writing header and lines together can fail. A typical mistake is submitting header and lines to the target system in one shot; if one line fails, the entire document must be redone. The stable practice is phased writing.
• Code mapping scattered across multiple Excel sheets. Many clients initially keep warehouse, vendor, and SKU mappings in scattered tables, making troubleshooting difficult. We later centralized them in Qeasy's mapping center with version traceability, making changes transparent at a glance.
• Full sync started before freezing the incremental point. A common pitfall—the business team is supplementing documents while integration runs full sync, resulting in mismatched numbers on both sides, leading to lengthy troubleshooting.
• Draft documents synced over. If unapproved documents from the source system are synced downstream, inbound amounts will be inflated or be unapproved. Be sure to add the approval status to the filter condition.
• Failure alerts ignored. After the strategy goes live, if the exception queue only piles up without handling, problems accumulate. It is recommended to configure DingTalk or WeChat Work alerts in Qeasy for immediate notification.

Applicable and Non-applicable Scenarios

Applicable: Retail, e-commerce, and wholesale supply chains where Kingdee Cloud Cosmos serves as the ERP master data source and Wangdiantong handles purchase inbound execution, requiring distributed transfer-in documents to be reliably converted into purchase inbound vouchers.

Not applicable: On-site operations requiring second-level real-time response, scenarios requiring two-way write-back with strict financial reconciliation requirements, and early stages where item master data governance has not yet been completed.