How to connect Power BI to Infor LN
Short answer
Connect Power BI to Infor LN through a read replica or warehouse, the Infor Data Lake, or ION API extracts, using an on-premises data gateway for private databases. Model items, orders and projects in a star schema, then add natural-language questions on top.
LN has its own reporting, but most manufacturers end up wanting Power BI for executive and cross-system views. The difficulty is not the visuals, it is the data: LN tables are named by package and module, carry a company suffix, store codes and enumerations that need decoding, and sit on a transactional database you do not want to hammer.
A good build separates the layers: extract from LN into something reporting-friendly, model it once, and let Power BI consume the model. Skipping that and pointing reports at live transactional tables works for a prototype and fails at scale.
The choices below depend on your database, whether you license the Infor Data Lake, and whether the site is on-prem or cloud.
What teams use it for
On-time delivery and supplier scorecards
Receipt dates versus promise dates by supplier and commodity, with rejection rates, giving buyers a basis for performance discussions.
Open order and backlog view
Sales order lines by promised date, status and value, with customer filters, for sales and operations meetings.
Inventory health
On-hand by warehouse, ageing and turns by item group, with slow-moving and excess highlighted.
Production schedule adherence
Planned versus actual completion by work centre and order, showing late orders and causes.
Project margin and cost tracking
Budget, committed and actual by project element for engineer-to-order work, with variance trends.
Cross-system margin analysis
Combine LN cost and sales with CRM or MES data in one model for margin by product and customer.
Infor LN integration surface
- Direct read from the database replica: SQL Server or Oracle (and DB2 in some estates) accessed with a read-only account; table names combine package, module and sequence with the company number.
- Infor Data Lake and Data Fabric, where licensed: replicated LN objects with SQL access, which isolates reporting load from production. Connectivity options depend on your Infor OS version.
- ION API and BOD extracts: suitable for moderate volumes and API-based refresh, less so for large history loads.
- Session exports and scheduled reports: a stopgap for small datasets that avoids touching the database but is hard to govern.
- Enumerations and domains: LN stores codes for statuses; you must map them to labels in your model or reports will show numbers.
How it works
- 1
Choose the source of truth
Prefer a replica, warehouse or Data Lake for reporting. Avoid pointing refreshes at the live transactional database during planning runs.
- 2
Extract the needed tables
Pick the core tables for items, orders, receipts, inventory and projects. Include the company number as a key everywhere.
- 3
Decode and conform
Map status enumerations to labels, standardise dates, units and currencies, and create dimensions for item, supplier, customer and calendar.
- 4
Set up the gateway
For an on-prem source install the on-premises data gateway on a server that reaches the database, secure its service account and register it in the Power BI service.
- 5
Model for performance
Build a star schema with fact tables for order lines, receipts and inventory snapshots, use incremental refresh for large facts and keep DAX measures simple.
- 6
Apply row-level security
Restrict by company, site or business unit so users see only their scope, and test it with real accounts.
- 7
Add natural-language questions
Enable Q&A on a clean semantic model with synonyms, or add a grounded assistant that queries the same model with governed definitions.
Example flows
Nightly supplier scorecard
Trigger: Scheduled refresh after the nightly extract.
- Extract loads purchase receipts and order lines to the warehouse.
- Model computes on-time and rejection measures by supplier.
- Power BI service refreshes through the gateway.
- Scorecard report updates for buyers before the morning meeting.
Result: Buyers review current supplier performance without asking IT for an export.
Ask a question of the model
Trigger: A manager types 'late orders by customer this month'.
- Q&A or a grounded assistant interprets the question using model definitions.
- It queries the semantic model, not the LN database.
- A visual returns with the filters shown.
- Manager pins it to a dashboard or drills in.
Result: Answers to ad hoc questions without building a new report each time.
Keeping data on-prem
Power BI Desktop and Report Server can work against an on-prem LN estate, and the Power BI service reaches private databases through the on-premises data gateway. If policy forbids sending data to the cloud service at all, Power BI Report Server keeps reports inside your network, with feature differences from the service. Natural-language questions on a fully private estate need a locally hosted model querying the same governed model, since hosted Q&A features depend on the cloud service.
Security checklist
- Use a read-only database account and keep gateway credentials in a controlled service account.
- Apply row-level security by company and business unit and test it with non-admin users.
- Exclude or mask sensitive fields such as customer pricing and controlled technical data from the model.
- Restrict workspace access and sharing, and review who can export data from reports.
- Audit refresh failures and unusual query loads on the replica.
Ways to get it done
Direct connection to a replica using the gateway
Effort: Days to weeks for first reports.
Single-site LN with modest data and a small reporting team.
Data warehouse or Infor Data Lake feeding Power BI
Effort: Weeks to months, with modelling upkeep.
Multi-site, multi-system reporting and larger history.
Partner-built model plus private natural-language layer, such as Netray
Effort: Scoped pilot of several weeks.
Teams wanting governed definitions and on-prem question answering.
Pitfalls
- !Reporting directly off live transactional tables slows planning runs and produces inconsistent numbers.
- !Forgetting the company number in relationships double counts or mixes entities.
- !Showing raw status codes instead of labels confuses users and erodes trust.
- !Full refreshes of large facts exhaust gateway capacity; use incremental refresh.
- !Different report authors defining 'on time' differently creates conflicting scorecards; define measures once in the model.
Frequently asked questions
Is there a standard Power BI connector for Infor LN?
We are not aware of a dedicated one that covers all setups, so check current options. Common routes are the underlying database connector on a replica, the Infor Data Lake where licensed, or API extracts, each with different effort and governance trade-offs.
Can we connect straight to the LN database?
Technically yes with a read-only account, ideally on a replica. It is quick for a prototype, but live transactional tables are hard to model and can load production. A warehouse or Data Lake layer is more sustainable for anything beyond small reports.
Do we need an on-premises data gateway?
If the data source is inside your network and you use the Power BI service, yes. The gateway lets the cloud service refresh from the private database without opening inbound ports. Report Server inside your network avoids the need but has fewer features.
Can users ask questions in plain English?
Yes. Power BI Q&A works best on a clean semantic model with synonyms. For richer, governed answers, including documents and record lookups, a separate assistant over the same model can help. Fully private sites need a locally hosted model for this.
How often can data refresh?
That depends on source, volume and licence. Nightly is common and easy; several times a day works with incremental refresh from a replica. Near real time usually needs event-driven extracts and is rarely needed for management reporting in manufacturing.
Want Power BI working on your Infor LN?
A free 30 minute assessment: we map the integration path for your version and setup, and where it should run.
Get a free assessment