Qeasy Cloud
Get Started

Integration Solution Overview: Yonyou BIP and Yonyou YS Inventory Sync

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

Scenario and Value

Three days after last year's Double 11, the finance team at a retail company was still reconciling in Excel: the Yonyou YS transit inventory ledger on one side, Yonyou BIP supply chain balances on the other. The numbers were off by hundreds of thousands, and only after business teams pushed back did they realize the closing quantities had not been synced into BIP on a daily basis. This was not an ERP error—it was the absence of an automated loop between the two systems.

In customer engagements we repeatedly see the same pattern: YS produces the ledger, BIP handles reconciliation and negative-balance checks. When the ledger fails to land by early morning, negative-balance alerts are delayed and reference price data is missing. This solution closes that gap by stably syncing the transit inventory ledger and inventory prices on a daily incremental basis, so business teams can use the data the next day without chasing the ERP.

Integration Architecture and Data Flow

The architecture is organized into three layers—source system, integration platform, target/cache—with the integration platform owning ETL, code mapping, incremental filtering, and retry logic:

┌────────────────────────────────────────────────────────┐
│  Yonyou YS (Inventory / Supply Chain)                    │
│  · Transit inventory ledger report (uniform/report/list) │
│  · Inventory price query interface                       │
└────────────────────────────────────────────────────────┘
                       │
                       ▼
┌────────────────────────────────────────────────────────┐
│  Integration Platform (Qeasy / 轻易云数据集成平台)         │
│  · Field mapping · Code mapping (material/warehouse/org/  │
│    unit) · Incremental filter · Full-volume fallback     │
│  · Retry · Dead-letter · Scheduling                      │
└────────────────────────────────────────────────────────┘
                       │
                       ▼
────────────────────────────────────────────────────────
│  Yonyou BIP (Supply Chain Master)                        │
│  · Inventory reconciliation · Negative-balance check ·   │
│    Price master data reference                            │
└────────────────────────────────────────────────────────┘

The data flow runs in two segments:

  1. Transit inventory ledger: YS emits the report; the integration platform pulls increments by document date (vouchdate), applies mappings for material, warehouse, organization, and unit, then writes into BIP's reconciliation/detail data, triggering the negative-balance check dimension navigation.
  2. Inventory price query: YS returns material price data; the integration platform caches it via a "write to no-op" target without persisting into BIP, serving as price master reference for downstream comparison and sync. The two strategies have no strong dependency and can run in parallel.

Interface List

Strategy IDData ObjectDirectionNotes
②Receive transit inventory ledgerYonyou YS → Integration Platform → Yonyou BIPSYNC, incremental by vouchdate, negative-balance check
〇Query inventory priceYonyou YS → Integration Platform (cache only)QUERY_ONLY, incremental by pubts, no write to target

Implementation Points

Phased scheduling. Both the ledger and price query run daily and should be staggered: ② transit inventory ledger at 02:00, 〇 inventory price starting at 02:13, to avoid contention on YS APIs. With sufficient resources, the two strategies can run in parallel since they have no mutual dependency.

Incremental fields and full-volume fallback. The ledger uses startvouchdate = last_sync_date and endvouchdate = CURRENT_DATE - 1 day for increments. endvouchdate is always set to the previous day to prevent same-day unposted documents from polluting reconciliation. For full-volume runs, pull startvouchdate back to the system go-live date and replay history. The price query filters by pubts > last_sync_time; if the interface does not support pubts, fall back to full volume. LAST_SYNC_DATE / LAST_SYNC_TIME are maintained by the integration platform and updated after each successful landing.

Centralized code mapping. Mapping for material, warehouse, organization, and unit is where this solution trips most often. The mapping tables must be loaded into the integration platform before scheduling goes live. Source fields productNumber, warehouse, org, and unit_code must each map one-to-one to BIP master codes. Refresh mapping tables whenever material master data is synced, to avoid dirty data where "inventory arrives but material cannot be found."

Retry and dead-letter handling. API timeouts and 5xx errors use exponential backoff at 30s/60s/120s, up to 3 attempts; 429 rate limiting waits 60s and retries up to 5 times. Business validation failures do not enter the retry queue—they go straight to dead-letter for human intervention. Single-record failures must support checkpoint resumption so one bad record does not block the whole batch. Trigger an immediate alert on 3 or more consecutive failures, a daily summary alert when the failure rate exceeds 5%, and another alert when data lag exceeds 2 scheduling cycles.

Privacy and credentials. Tenant IDs, organization codes, API credentials, and connection strings must come from environment variables or a secrets manager—never hard-coded into solution docs or scripts. YS APIs are recommended to authenticate via OAuth2 or API Key.

Best Practices and Pitfall Reviews

  1. endvouchdate always means T-1. In one manufacturing project, we initially set endvouchdate to CURRENT_DATE, and unposted same-day documents slipped into BIP, generating false negative-balance alerts. The safe pattern is to fix it as DATE_SUB(CURRENT_DATE, INTERVAL 1 DAY), so the reconciliation view is naturally one day behind, which business users already expect.
  2. Phased loading of code mapping. Material master data syncs first; mapping tables are loaded per material and refreshed periodically. Only after that do we kick off the transit ledger and price query. On a customer site we used this "master data first, inventory second" rhythm in Qeasy (轻易云), lifting first-time reconciliation success from below 60% to above 95%—the key is isolating inconsistent master data upstream of the inventory sync.
  3. QUERY_ONLY still needs an exit. Inventory price is query-only, but if it just lingers in the integration platform cache it quickly becomes unmaintained dirty data. Configure an explicit "write to no-op" target and maintain an incremental cursor by pubts, so the price master stably feeds downstream comparison and reporting.
  4. Lock dimension navigation to negative-balance check. The dimension navigation scheme on the ledger strategy is fixed to "negative-balance check." Do not let downstream change it arbitrarily, otherwise the negative-balance alert chain gets fragmented.

When to Use Qeasy

In inventory scenarios with many organizations and warehouses where reconciliation views are scattered, Qeasy (轻易云数据集成平台) turns code mapping, incremental cursors, retry, and scheduling orchestration into visual configuration. The "master data first, inventory second" rhythm can be reused directly, and a daily reconciliation loop can be brought online within a single iteration—instead of writing yet another batch of scripts.

Original content. Please credit the source when reposting: https://www.qeasy.cloud/insights/solutions/sol-bip-ys-1738

Comments