Qeasy Cloud
Get Started

Practical Guide to Syncing Other-Outbound Stock Orders from WMS to Kingdee Cloud Galaxy

· 尹春锐· Integration Solutions· 24 views· 5 min read
WDTKingdee Cloud供应链集成Inventory Sync轻易云实战教程

What This Strategy Solves

In retail and distribution businesses, the front end for shipping is usually a WMS-style platform, while the financial and inventory ledger of record sits in an ERP system such as Kingdee Cloud Galaxy. Both sides carry an Other-Outbound order document, but they serve different purposes: the front end cares about picking, wave planning, and logistics, while the back end cares about cost, inventory accounting, and the general ledger. In one real engagement, the customer's recurring pain point was simple — every outbound on the front end had to show up in the back-office ledger promptly, with no gaps across days or organizations. The strategy that syncs Other-Outbound orders from the WMS to the ERP is built precisely for this scenario.

Data Flow and Field Mapping

The flow is unidirectional: source system (WMS enterprise gateway) → Qeasy data integration platform (middleware) → target system (Kingdee Cloud Galaxy). The middleware does not persist business data; it only transforms, maps, cleans, and enriches.

Key field comparison:

Business MeaningSourceTargetNotes
Document numberOutbound order numberDocument number + source document numberUse source number as target source doc number for traceability
WarehouseWarehouse codeWarehouse code (under organization)Warehouses must be set up on both sides with consistent or mappable codes
Item codeProduct code / merchant codeMaterial codeA common pitfall — see below
QuantityOutbound quantityActual issue quantityWatch for unit conversion
Posting dateAudit timestampBusiness dateCross-midnight documents should use the business date
Document typeOutbound typeBusiness type (Other Outbound)Fixed value mapping
RemarksRemarksRemarksPass through as-is

How to Configure on Qeasy

We use the Qeasy data integration platform to orchestrate the entire flow. Typical configuration falls into four parts.

1. Source connector. Select the WMS enterprise gateway adapter, grant tenant authorization, and pull the Other-Outbound order interface. For the incremental starting point, place it safely after the most recent stable audit timestamp rather than at midnight on go-live day. This avoids having legacy dirty data flood in at launch.

2. Target writer. Select the Kingdee Cloud Galaxy adapter. Organization isolation is a hard requirement on the target side, so the organization (corresponding to the customer's financial organization) should be hard-coded in the connection, not parameterized. Shared connections with parameter substitution frequently cause cross-organization contamination.

3. Centralized code mapping. Warehouses, item codes, and organizations are reused dimensions across many strategies. The pattern we use is to maintain a code-mapping table centrally in the platform rather than scattering mappings across each strategy. New strategies simply reference the existing map, and changes propagate everywhere.

4. Phased header and body rollout. Phase one pushes only the header and base quantities to validate end-to-end connectivity. Phase two adds unit price, batch, and remarks — fields that more often cause trouble. This is a common Qeasy customer pattern: header-first, body-second staging.

Implementation Steps

Step 1 — Environment and connection. Set up both connections in Qeasy and run connectivity self-tests. The most common stumble here is API throttling on the target side; keep the call rate under 1–2 requests per second initially and only relax it after observation.

Step 2 — Field mapping and trial run. Load the field mappings into the platform and trial-run on a few dozen sample records. Focus verification on business date, warehouse, and organization — if rows are not crossing rows or organizations, you can proceed.

Step 3 — One-shot full load. Pick a low-traffic business window and run a historical full load to push existing records. A human must watch this run; failed records must be cleared immediately or they snowball into the next attempt.

Step 4 — Schedule and increment. Inventory documents should not run on long intervals (such as once a day). A 10–30 minute polling interval for incremental sync is recommended. Once the incremental starting point is fixed, do not move it. This is the essence of the dual-track pattern — a full load runs once at cutover, while daily operation runs entirely on the incremental track.

Step 5 — Monitoring and alerting. The monitoring dashboards built into Qeasy are sufficient. The important configuration is failure retry counts, upper-bound failure alerts, and a daily reconciliation task comparing document counts on both sides.

Lessons from the Field

  1. Item code misalignment. This is the number-one pitfall. Merchant codes on the source side and material codes on the target side often differ. Pushing hard results in rejection or stray codes. The safe approach is to look up a central map in the middleware; if no mapping exists, route to manual remediation.

  2. Business date on cross-midnight documents. A document audited at 1 a.m. — does it belong to today or tomorrow? Warehousing and finance often disagree. Decide this up front. We recommend always using the natural day of the audit timestamp on the warehouse side; this avoids endless reconciliation debates later.

  3. Cross-organization contamination. The target ERP isolates data by organization. A user can see warehouses from another organization, but if the write picks the wrong organization, the record lands in someone else's books. Cleaning that up is painful. Do not parameterize the organization in configuration.

  4. Duplicate rows from retry storms. Under network instability it is hard to distinguish a true timeout from data that was actually written. A common Qeasy pattern is to enable idempotency keys — using the source document number as the unique key — and to query before retrying; if the record exists, skip it.

  5. Full-load and increment colliding. We once saw a full-load job and the increment job run simultaneously and push the same document twice. After that, the rule became explicit: full load runs once on cutover day and its switch is immediately turned off; increment runs 24/7, paused briefly during the full-load window.

When to Use and When Not

Use it when the front-end WMS and back-office ERP are separated and warehouse outbounds must be reflected in the financial inventory ledger in near real-time, and when the document structure is stable and does not need heavy recomposition. Do not use it when the middleware must perform complex business validation or restructuring, nor for very high volumes with sub-second latency requirements — those scenarios are better served by direct connection over a message bus rather than scheduled sync.

Original content. Please credit the source when reposting: https://www.qeasy.cloud/insights/solutions/strat-wdt-kingdee-cloud-9833-n58648494-829c3116

Comments