Oracle Fusion Cloud ERP Guide
Setting up the Oracle Fusion Cloud ERP Integration
To access your users' data from Oracle Fusion Cloud ERP, you need to enable the Oracle Fusion Cloud ERP integration in the Platforms Page of the Rutter dashboard.
Oracle Fusion Cloud ERP does not require any additional developer app setup.
User Authentication with Oracle Fusion Cloud ERP
A Rutter Connection is created when your end user authenticates their Oracle Fusion Cloud ERP account via Rutter Link. This in turn allows you to access their account data via the Rutter API.
Oracle Fusion Cloud ERP uses HTTP Basic authentication. The end user is prompted to enter the following fields in Rutter Link:
- Instance URL — the base URL of their Oracle Fusion Cloud Applications tenant, including the protocol but no trailing path (e.g.
https://your-tenant.fa.us2.oraclecloud.com) - Username — the username of an integration user with at least read access to the relevant data
- Password — the password for that integration user
Accounts
Accounts are read-only through GET /accounting/accounts and GET /accounting/accounts/:id. Rutter reads Oracle's /fscmRestApi/resources/11.13.18.05/accountCombinationsLOV collection and its /{_CODE_COMBINATION_ID} resource. Account creation, updates, and deletion are not supported.
- Each account represents a General Ledger account combination.
platform_idis the exact Oracle_CODE_COMBINATION_ID(CCID), not a natural-account segment or full posting code. Rutter'sidis a separate persisted identifier. _CHART_OF_ACCOUNTS_IDremains available as raw chart context inplatform_data. Scans do not select a single chart, but this does not establish that all accounts or charts visible in Oracle's setup screens are accessible to the integration user, or that posting codes resolve to these combinations.name,nominal_code,currency_code,balance,parent_id,created_at,updated_at, andplatform_urlare unsupported and null; legacy API versions retain their existing empty-stringupdated_atconvention. Rutter does not construct a name or posting code from CCID, chart ID, or flexfield order. Subsidiary associations are empty.- Oracle's A/L/E/O/R account types map to broad asset/liability/expense/equity/ income categories. Unrecognized types remain unknown; an asset's subtype is unknown. Posting, summary, and financial-category flags do not determine status.
Status is a snapshot computed during fetch using the UTC calendar date and Oracle's EnabledFlag, StartDateActive, and EndDateActive. Disabled accounts or accounts whose end date has passed are inactive. Otherwise, a future start date is pending, and the account is active once its start date is reached. Both boundary dates are inclusive; null or absent dates are open-ended. Reads do not recompute status: it can remain stale across a date boundary until the next successful scan. A status-boundary change updates Rutter's persisted snapshot even when Oracle's fields have not changed. This is a Rutter UTC policy, not a claim about the tenant's local timezone. The internal status marker is not exposed in platform_data; native fields and unknown flexfields are retained.
Full and incremental (side) refreshes rescan the collection because no modification timestamp or modified-since filter is established. Rutter requests up to 500 rows per page using the existing Oracle Fusion limit; Oracle may return smaller pages. Oracle scans continue by CCID keyset: the first query uses q=_CODE_COMBINATION_ID>0, and each next query uses the last returned CCID with orderBy=_CODE_COMBINATION_ID:asc and no offset. Internal Oracle fetch checkpoints contain only the exact last CCID; older offset checkpoints are rejected and must restart the scan, not converted ambiguously. Public Rutter API cursors are unchanged. Malformed identities, rows, enabled flags, date ranges, or pagination fail the fetch rather than silently skipping data; provider errors propagate. This avoids offset-shift skips of surviving later IDs, but is not a point-in-time snapshot: records inserted or changed behind the cursor need a later rescan.
Each scan costs roughly ceil(record_count / effective_page_size) list requests, plus retries or separate ID lookups. Tenant record counts, effective page caps, latency, appropriate refresh frequency, completeness, and API/UI parity still require live validation. The requested page size is not evidence of tenant capacity. This standalone Accounts implementation does not establish Bills or payment posting-code linkage or complete the live-validation release gates.
Item Catalog
The items GET endpoint exports Oracle Product Hub catalog items from itemsV2. The integration user needs read access to Product Hub items in the inventory organizations being exported; General Ledger access alone is not sufficient.
- Each inventory-organization assignment is a separate item. Its
platform_idcombines Oracle'sItemIdandOrganizationId, separated by__. The same item code can appear in multiple organizations with different purchase defaults. nameand the bill/invoice descriptions use the short item description;codeuses the item number.bill_item.unit_priceis the Purchasing purchase default (ListPrice). Missing prices remain null.invoice_item.unit_priceandquantity_on_handare null because sales pricing and on-hand balances are not supplied by this catalog integration.subsidiariesis empty: an inventory organization is not a legal subsidiary, and this integration does not yet resolve that relationship.- Oracle's Active and Inactive status labels map to active and archived. Other configured labels map to unknown. Inventory items are distinguished using
InventoryItemFlag; non-inventory does not necessarily mean service. - Full and incremental sync use the last-update timestamp. Receivables memo-line-only services and item writes are not included.
Tax Rates
Oracle Fusion tax rates are read through GET /accounting/tax_rates. The integration user needs read access to Oracle's taxRatesLOV REST resource.
- Each exported record represents a tax-rate period.
platform_idis Oracle'sTaxRateId, not its tax code. Multiple records can share a code when the rate changes over time. - Historical and future periods are retained.
statusreflects the last successful sync, using the UTC date at fetch time and the period's effective dates, including the end date. Future and expired periods areinactive; a missing start date producesunknownunless the end date already establishes that the period has expired. A period crossing its start or end date is reflected after the next successful sync, not automatically when a GET is made. total_tax_rateis the nominal percentage, such as3.5for 3.5%, from Oracle'sPERCENTAGErate type. Recovery rates are excluded: a recovery percentage describes how much tax is recoverable, not the transaction tax rate. Quantity amounts are not converted into percentages.effective_tax_rateisnullbecause this resource does not establish compounding. Components and subsidiary links are empty, and sales/purchase-specific rates arenull.- Oracle does not expose creation or modification timestamps on this resource, so
created_atandupdated_atarenull. Both historical and incremental syncs scan the complete resource; effective dates are not used as modification timestamps. - Rutter continues scans by
TaxRateIdrather than offset: Oracle's LOV offset pagination can report exhaustion at 5,000 rows even when more rates exist. platform_data.__rutter_oraclefusion_effective_period_statusis Rutter-derived metadata, not an Oracle field. It stores the status calculated at fetch time so the returned status and indexed status filters agree. The marker changes only when status changes; it does not store the current date.
Oracle Fusion tax-rate support is read-only. This integration does not add support for creating tax rates.
Subsidiaries
Rutter's subsidiaries endpoint exports the legal entities accessible to the integration user, using Oracle Fusion's legalEntitiesLOV resource. platform_id is the Oracle LegalEntityId. Business units and ledgers are different organizations and do not share this identity.
The endpoint supports paginated reads. Incremental syncs reread the legal-entity list because Oracle does not expose modification timestamps on this resource. Status is active when the current UTC date falls within the entity's start and end dates, including both boundary dates. An absent boundary is open-ended. Status changes are reflected on the next sync.
The resource does not expose country, currency, parent-subsidiary hierarchy, elimination status, or creation/update timestamps. Rutter returns an empty country, null currency_code, parent_id, created_at, and updated_at, and defaults elimination to false. These defaults do not establish the corresponding values in Oracle's setup screens; in particular, a country shown in the Oracle UI is not included in this API resource.
Transaction associations must use an explicit legal-entity reference or an unambiguous legal-entity lookup. A business-unit or ledger ID must not be treated as a subsidiary ID.
Vendors
Vendors are read-only through GET /accounting/vendors and GET /accounting/vendors/:id. The integration user needs supplier-read privileges; General Ledger access alone is insufficient. Vendor creation, updates, and deletion are not supported.
platform_idis Oracle'sSupplierId, not itsSupplierNumber. Both vendor and display names useSupplier, notAlias.- Contact name, email, and phone come from the lowest-ID active contact, falling back to the lowest-ID contact when none are active. The administrative-contact flag is not treated as primary. Phone components are joined as
+country area number ext. extensionwhen present; this is not E.164 normalization. - All addresses, including inactive ones, are exported in ID order. Remit-to addresses are billing addresses; other purposes are unknown, not inferred as shipping.
- Currencies are the sorted, unique ISO codes from the preferred functional currency and every site's invoice and payment currencies, including inactive sites.
- Payment terms, subsidiaries, bank account numbers, and business registration numbers are not mapped. Procurement business units are not legal entities, and
RegistryIdis an internal identifier, not a business registration number. Full hydrated supplier data remains available inplatform_data.
Supplier refreshes rescan all records, including side refreshes: LastUpdateDate filtering returns HTTP 400, and child-only changes may not update the supplier timestamp. Addresses, contacts, and sites are independently paginated beyond Oracle's default 25-row page. Site visibility is business-unit scoped; UI and API site counts can differ, and parity is not established.