Supplier Master Sync in Practice: A Steady Rollout from Kingdee Cloud to Wangdiantong QM
What This Strategy Solves
In a real engagement with a retail client, the procurement team reported a recurring issue: a supplier was created in the ERP, but by the time the buyer tried to place a purchase order in the front-end system, the supplier could not be found. Staff ended up re-creating the supplier manually, and within weeks the codes on both sides drifted apart, throwing off reconciliation. This strategy pushes supplier master data on a schedule so that one master serves many systems and codes do not drift.
Data Flow and Field Mapping
The flow is one-way: source system (back-office ERP, here represented by Kingdee Cloud) → Qeasy integration platform → target system (front-end business system, here represented by Wangdiantong QM). The integration layer only transports and maps; it does not run business logic.
Key field mapping (representative, not exhaustive):
| Meaning | Source field | Target field | Notes |
|---|---|---|---|
| Supplier code | FNumber | provider_no | Primary key, centralized mapping |
| Supplier name | FName | provider_name | Required |
| Create org | FCreateOrgId.FNumber | org_no (target-side) | Org mapping required |
| Use org | FUseOrgId.FNumber | org_no | Explode per org |
| Remark | FDescription | remark | Optional |
The source side uses executeBillQuery to pull FSupplierId, FNumber, FName, and org fields. The target side uses an execute-style API such as provider.create. FSupplierId acts as the source-side unique key; the id returned from the target is written back as the anchor for incremental dedup.
How to Configure on Qeasy
Source strategy (QUERY direction): pick executeBillQuery, POST, select FSupplierId, FNumber, FName, and the org fields; turn on idCheck and dedup by FSupplierId.
Target strategy (EXECUTE direction): pick the supplier-create API, POST, and reference source-side values with template syntax such as {{FNumber}} and {{FName}}.
On Qeasy, keep code mapping and org mapping in one centralized mapping table rather than scattering them across strategies. When a new business org is added or coding rules change, only one place is edited.
Implementation Steps
We recommend three phases.
Phase 1 — incremental seed. First check whether a last-modified timestamp is available in the source; if yes, use it as the incremental seed, otherwise fall back to full sync. Qeasy supports a dual-track pattern: run incremental during the week and trigger a full reconciliation on a fixed cadence.
Phase 2 — full trigger. During cutover or whenever code drift is suspected, manually trigger a full sync. Page by FSupplierId ascending on the source side, upsert record by record on the target side. After the run, the numbers on both sides must match; any delta becomes the starting point of investigation.
Phase 3 — scheduling cadence. The source strategy can run every 10 minutes (offset from the top of the hour), the target strategy every 20 minutes. Staggering gives the target side enough time to write back the id before the next incremental round. The active window should align with business hours, for example 07:00 to 22:00.
Lessons from the Field
Pitfall 1: code mapping not centralized. A typical mistake is to encode the mapping inside every strategy, so a single change touches a dozen places. The safe approach is to maintain one global mapping table on Qeasy and have all strategies reference it.
Pitfall 2: org fields not exploded. One source-side supplier may map to multiple use orgs; if only the first is taken, branches are dropped on the target. The safe approach is to explode multi-org values into multiple rows on the source side and write one row at a time on the target.
Pitfall 3: wrong incremental seed. Using a created-time seed causes suppliers whose names changed historically to be missed. The safe approach is to use last-modified time and, before the first run, align a baseline with the source system.
Pitfall 4: idCheck left off. Without dedup, network retries create duplicates. The safe approach is to keep idCheck on at the source and rely on the written-back id for idempotency at the target.
Pitfall 5: no reconciliation after a full run. A typical mistake is treating the run as finished and discovering a few missing rows only later. The safe approach is to trigger an automatic reconciliation right after a full run and log any diff.
When to Use and When Not to
Use it when a multi-org enterprise keeps supplier master data in the back-office ERP and the front-end business system needs the same master per org or per store, with a unified coding rule and clean org mapping. Do not use it when supplier master data is owned by the front-end system and the ERP only receives, or when the two sides have heavy business-field divergence that requires custom development.