Qeasy Cloud
Get Started

Practical Guide: Syncing Other Inbound Orders from Kingdee to Outsourced Inbound Orders in Wangdiantong on Qeasy

· 吕修远· Integration Solutions· 14 views· 4 min read

What This Strategy Solves

In one outsourced-processing engagement, a manufacturing client generated dozens of "Other Inbound Orders" in Kingdee Cosmos every day. Those orders had to be pushed into Wangdiantong as "Outsourced Inbound Orders" under the outsourced warehouse. The two systems did not talk to each other directly, so finance and warehouse staff reconciled by hand, which was error-prone and slow. This strategy turns that intermittent manual flow into an automated, traceable, scheduled sync.

Data Flow and Field Mapping

Direction: Kingdee Cosmos (system B) → Qeasy middleware layer → Wangdiantong (system A, Outsourced Inbound Order, 3000 Cloud Warehouse).

We typically maintain a central mapping table in the middleware layer to manage code and header/body differences in one place. Below are the key fields we usually confirm in real projects.

DimensionSource (Kingdee Other Inbound)MiddlewareTarget (Wangdiantong Outsourced Inbound)
Document No.FBillNoPass-throughouter_no
Document DateFDateyyyy-MM-ddorder_time
Warehouse CodeFStockId.FNumberWarehouse mapwarehouse_no (outsourced)
Supplier / VendorFSupplierId.FNumberCentral mappingcustomer_no
Item CodeFMaterialId.FNumberCode mapgoods_no
QuantityFQtyUOM conversionnum
RemarksFNoteTrim & cleanremark

Note that Wangdiantong's Outsourced Inbound Order usually requires an explicit outsourced-warehouse flag. Centralized code mapping is the rule we set with clients; later vendor or warehouse changes happen in one place, avoiding the blast radius of hard-coded mappings.

How to Configure on Qeasy

On the Qeasy data integration platform, the strategy is split into three objects: source extraction, transformation, and target write.

Source side: pull "Other Inbound Orders" from Kingdee Cosmos via its query interface, filtered by modification time for the initial increment.

Transformation layer: in Qeasy's mapping canvas we do the following.

  • Code mapping: suppliers, items, and warehouses go through a central mapping table.
  • Header and body staged: persist the header first, then write body rows by internal id, avoiding the missing-row pitfall of one-shot writes.
  • Status control: only approved documents flow downstream; drafts are filtered out.

Target side: call the Wangdiantong Outsourced Inbound Order API. The warehouse-type field must carry the outsourced warehouse code, otherwise the order will be treated as a regular inbound. That is a classic failure point.

Qeasy routes failed records into an exception queue. In customer projects we typically add a retry-with-email-alert fallback.

Implementation Steps

We recommend a three-step rollout.

Step 1 — Increment baseline. Pull recent approved Kingdee Other Inbound Orders ordered by FDate + FBillNo. This is a one-shot manual run to validate field alignment end to end.

Step 2 — Full backfill. On first go-live, run a full backfill to fill historical documents. Qeasy's full-load task supports pagination and checkpoint resume, preventing gaps.

Step 3 — Normal scheduling. Move into the dual-track mode that we commonly see among Qeasy customers.

  • Increment: every 10–15 minutes, filtered by FModifyDate.
  • Full: nightly during off-peak, used as a reconciliation sweep; differences go to a diff table.

This dual-track pattern keeps latency low during the day and reconciles at night.

Pitfalls We Have Hit

  1. Warehouse type defaults to a regular warehouse. Outsourced orders must carry the outsourced warehouse code explicitly; otherwise downstream picking assigns the wrong slot. This is the most common failure.

  2. Header/body write order. Writing body rows before the header creates orphan child rows. Staging header first, then body, is the safe pattern.

  3. Wrong increment field. Using document date as the increment cursor misses edits, since document date rarely changes. We consistently use FModifyDate as the increment baseline.

  4. Mapping scattered in scripts. Adding a new vendor meant editing a dozen scripts. After centralizing mappings, that drops to a single edit.

  5. Status not gated. Drafts leaked into the target and triggered null-field errors. Adding "status = approved" in the transformation layer is the safe default.

When This Applies and When It Does Not

Applies: manufacturing or distribution clients where Kingdee Cosmos and Wangdiantong outsourced-warehouse boundaries are clear, codes are relatively stable, and both systems expose APIs. Does not apply: cross-entity projects, projects that need secondary cost accounting, or projects where coding rules are frequently reshaped — those need a code-governance phase before automation is realistic.

Original content. Please credit the source when reposting: https://www.qeasy.cloud/insights/solutions/strat-wdt-kingdee-cloud-0874-3000-a825d674

Comments