Does the integration require an on-premises agent to connect to SAP ECC?

Yes. SAP ECC is on-premises only and exposes data via RFC/BAPI and OData, which require an on-premises agent (SAP .NET Connector or Java Connector) running on the customer's network. ml-connector routes all SAP calls through this agent. The Looker side is cloud-native and requires only OAuth2 credentials.

What happens when Looker's OAuth2 token expires after one hour?

ml-connector tracks token expiration and automatically exchanges the OAuth2 client credentials for a fresh token before the old one expires. If a call returns 401 Unauthorized, ml-connector refreshes the token immediately and retries the query, so token expiration is transparent to the integration.

How are SAP GL accounts and cost centers mapped to Looker query filters?

Master data polling from SAP ECC extracts GL account numbers and cost center codes, which ml-connector maps to the corresponding Looker dimension names in your warehouse model. During validation queries, ml-connector uses these mapped names in the query filter so results are aggregated on the correct warehouse columns and can be compared back to SAP totals.

SAP ECCLooker

SAP ECC and Looker integration

SAP ECC runs finance and procurement. Looker models and analyzes that data for business intelligence. Connecting the two ensures your financial reporting and data models stay in agreement. Master data such as vendors, customers, and GL accounts defined in SAP ECC are validated against Looker's modeled warehouse, and financial metrics from ECC can be cross-checked against Looker queries to catch data drift or transformation errors. ml-connector handles SAP's on-premises connectivity requirements and Looker's OAuth2 authentication, running the validation on a schedule you control.

How SAP ECC works

SAP ECC exposes vendors, customers, materials, GL accounts, cost centers, purchase orders, and accounting documents through RFC/BAPI function modules via an on-premises agent (SAP .NET Connector or Java Connector), OData v2 REST via SAP Gateway at http://<host>:<port>/sap/opu/odata/<namespace>/<service>/, or SOAP web services. Authentication is HTTP Basic Auth for OData and IDoc, or Basic Auth via RFC for BAPI calls. SAP ECC has no native webhook system, so data is read by polling on demand or on a scheduled interval. Key entities include BAPI_VENDOR_GETLIST, BAPI_VENDOR_GETDETAIL, BAPI_CUSTOMER_GETLIST, BAPI_GL_ACC_GETLIST, and RFC_READ_TABLE for custom table queries.

How Looker works

Looker is a business intelligence platform that models warehouse data and exposes it through REST APIs at https://<instance>.cloud.looker.com/api/4.0/. Authentication uses OAuth2 client credentials, exchanged via POST /api/4.0/login for a bearer token valid for one hour. Key entities include queries, looks, dashboards, users, groups, and folders. Looker does not natively store ERP data; instead it models data from connected warehouses. Pull-based querying via REST API is the primary integration pattern, as Looker has no inbound webhooks or real-time change events. Scheduled Plans can push query results outbound to webhook, email, S3, or SFTP.

What moves between them

The main flow runs from SAP ECC into Looker. Master data such as vendor lists, customer hierarchies, and GL account definitions are polled from SAP ECC and used to query corresponding Looker models for validation. Financial metrics, GL postings, and cost center allocations extracted from SAP ECC are cross-checked against Looker query results to verify that warehouse transformations are accurate and no data drift has occurred. Looker query results are read-only, so ml-connector extracts data for analysis but does not write back into Looker's modeled data.

How ml-connector handles it

ml-connector stores the SAP HTTP Basic Auth credentials and manages the on-premises agent routing required to reach RFC and OData endpoints on the customer's SAP system. For Looker, it exchanges OAuth2 client credentials for a bearer token on each session start and refreshes the token when a call returns 401, since Looker tokens expire after one hour. Master data polling from SAP ECC is scheduled on an interval you control, and ml-connector maps SAP GL account numbers and cost center codes to Looker dimension names so queries can filter and aggregate on the correct warehouse columns. Because both SAP ECC and Looker are pull-only on the ECC side, polling happens on a timetable aligned to your closing calendar. Every record carry a full audit trail, and failed validations are logged so data drift can be investigated before it impacts reporting.

A real-world example

A mid-sized manufacturing company runs SAP ECC for procurement and general ledger, and loads data nightly into a Snowflake warehouse for analytics via Looker. Before the integration, the finance and BI teams ran manual spot checks once per month to verify that Looker's vendor and GL account dimensions matched SAP's master data, which often surfaced transformation errors caught too late in the close process. With SAP ECC and Looker connected, ml-connector validates vendor lists and GL accounts every night after the warehouse load, flagging any dimension mismatches or missing records in Looker so the BI team can correct the warehouse pipeline before month-end close starts. Financial metrics are cross-checked automatically, and the manual validation step is eliminated.

What you can do

Poll vendor, customer, and GL account master data from SAP ECC and validate it against Looker's modeled dimensions.
Cross-check financial metrics and GL posting totals extracted from SAP ECC against Looker query results to detect data drift.
Authenticate SAP ECC via HTTP Basic Auth with on-premises agent routing, and Looker via OAuth2 client credentials with automatic token refresh.
Map SAP cost centers and GL accounts to Looker dimension names so data validation queries use the correct warehouse columns.
Schedule master data validation on a timetable aligned to your month-end close cycle, with a full audit trail on every check.

Questions

Does the integration require an on-premises agent to connect to SAP ECC?: Yes. SAP ECC is on-premises only and exposes data via RFC/BAPI and OData, which require an on-premises agent (SAP .NET Connector or Java Connector) running on the customer's network. ml-connector routes all SAP calls through this agent. The Looker side is cloud-native and requires only OAuth2 credentials.
What happens when Looker's OAuth2 token expires after one hour?: ml-connector tracks token expiration and automatically exchanges the OAuth2 client credentials for a fresh token before the old one expires. If a call returns 401 Unauthorized, ml-connector refreshes the token immediately and retries the query, so token expiration is transparent to the integration.
How are SAP GL accounts and cost centers mapped to Looker query filters?: Master data polling from SAP ECC extracts GL account numbers and cost center codes, which ml-connector maps to the corresponding Looker dimension names in your warehouse model. During validation queries, ml-connector uses these mapped names in the query filter so results are aggregated on the correct warehouse columns and can be compared back to SAP totals.

Connect SAP ECC and Looker

Free to use. Add your credentials, ping your real systems, and see if we fit.

Get started