›› Free Data Maturity Scorecard — Discover your analytics score in 3 minutes MY SCORECARD →
›› Free Fabric Capacity Estimator — Model your CU consumption before you upgrade ESTIMATE CAPACITY →
Cloud Data Platform Migration

OBIEE to Fabric Migration With the RPD Rebuilt as a Semantic Model, Not Just Dashboards Redrawn

An OBIEE or Oracle Analytics Server estate is more than its dashboards. The business logic lives in the RPD repository: physical, business model and presentation layers, calculations, hierarchies, variables and security. Our OBIEE to Fabric migration inventories that estate, rebuilds the model in Power BI on Microsoft Fabric, moves Answers and BI Publisher reporting across and proves parity before anyone switches.

RPD, Answers, dashboard and BI Publisher inventory first, so unused content is retired, not rebuilt
RPD logic rebuilt as a governed star-schema semantic model with tested DAX
Application roles and data filters mapped to row-level security
Documented handover and training for your Oracle BI developers
Migration at a Glance
3
RPD layers (physical, business model and mapping, presentation) rebuilt as one governed semantic model
3
Fabric semantic model modes: Import, DirectQuery and Direct Lake
4
Phases: discover, pilot, migrate in waves, cut over
Free
Oracle BI estate assessment to scope the work
Technologies Oracle BI (OBIEE) Oracle Analytics Server RPD Repository Answers BI Publisher Oracle Database Microsoft Fabric OneLake Power BI DAX Dataflows Gen2 Microsoft Entra ID
What We Cover

OBIEE to Fabric Migration: Rebuilding the RPD, Not Redrawing Dashboards

In Oracle BI, the dashboards are the visible tip. Beneath them sit the RPD, where physical tables are joined, a business model defines hierarchies, measures and aggregation rules, and a presentation layer exposes subject areas. Variables, initialization blocks, application roles and data filters add behaviour on top. An OBIEE to Fabric migration has to carry that logic across, not just recreate the charts.

The target is different in kind. A Power BI semantic model replaces the business model layer, DAX measures replace calculations written in logical SQL, and Power BI reports and paginated reports replace Answers analyses, dashboards and BI Publisher output. We resolve the model first, land the Oracle data in OneLake, then rebuild reports on top. Microsoft's migration guidance recommends focusing on the business questions a report answers, not copying every object exactly.

We decide per item. Content people use is rebuilt, near-duplicates are merged, and content nobody opens is retired. Support dates differ by Oracle BI release, so we do not quote one here. Oracle publishes them for OBIEE in My Oracle Support Doc ID 1664916.1 and for Oracle Analytics Server in Doc ID 2775561.1. Confirm yours with Oracle, and we will plan around it.

Challenges We Solve
"Nobody can tell us which analyses and dashboards are still used, or who owns them.
"Our key measures are defined in the RPD and only one or two people can safely change it.
"Our Oracle BI version is ageing and we are not sure what Oracle will support, or for how long.
"Application roles and data filters control who sees what, and we cannot lose that in a move.
What We Deliver

Six Core Components of Every OBIEE to Fabric Migration

From inventory to parity sign-off: everything needed to move Oracle BI and Oracle Analytics Server reporting to Power BI on Microsoft Fabric.

01
Oracle BI Estate Discovery & Triage

We list every RPD and subject area, Answers analysis, dashboard, BI Publisher report, agent and prompt, then check usage and owners. Each item is marked retire, consolidate or rebuild, with a complexity rating, before any build starts.

RPD, subject area and report inventory
Usage and owner audit
Retire, merge or rebuild decisions
02
RPD to Semantic Model Design on Fabric

We rebuild the business model layer as a star-schema semantic model: logical tables become fact and dimension tables, hierarchies and measures are redefined once, and the presentation layer becomes display folders and perspectives. We choose Import, DirectQuery or Direct Lake per subject area.

Business model to star schema
Hierarchies and measures defined once
Storage mode decision per model
03
Logical SQL & Calculations to DAX

We translate time-series functions, level-based measures, FILTER expressions, ranking, parent-child hierarchies and variables into DAX measures and model features, then test results at the original grain. Some logic moves upstream into Fabric transformations instead.

Time-series functions to time intelligence
Level-based measures to CALCULATE patterns
Variables to measures, parameters and RLS
04
Oracle Data into OneLake

We land the Oracle sources your RPD reads as Delta tables in OneLake. Where the source database fits Microsoft's requirements, mirroring can replicate Oracle tables with little custom work. Otherwise we use pipelines, copy jobs or Dataflows Gen2 with the Oracle connector through an on-premises data gateway where needed.

Mirroring where the source qualifies
Pipelines, copy jobs and Dataflows Gen2
Refresh schedules and credential setup
05
Answers, Dashboards & BI Publisher Rebuild

We rebuild priority analyses and dashboards as Power BI reports, mapping prompts to slicers and navigation to bookmarks and drill-through. Pixel-perfect BI Publisher output is rebuilt as paginated reports, and agents and scheduled delivery are replaced with subscriptions and alerts per report.

Report layouts focused on business questions
Paginated reports for operational output
Subscriptions and alerts where needed
06
Security, Governance & Parity Validation

We map application roles and data filters to workspaces, apps, roles and row-level security, then run a parity framework with report owner sign-off before cutover and set up usage monitoring afterwards.

Application roles to row-level security
Workspace and app access design
Totals, grain and filter parity checks
Where Most Migrations Go Wrong

RPD Logic to DAX: How We Translate Oracle BI Calculations

In the RPD, measures are defined once in the business model and Oracle BI generates the SQL. In Power BI, a DAX measure responds to the filter context of the report. So the translation is about controlling filter context, and every pattern is tested against the Oracle result. Illustrative patterns only: the right DAX depends on your model.

Oracle BI patternTime seriesAGO("Sales"."Revenue", "Sales"."Time"."Year", 1)
Typical Power BI approachCALCULATE ( [Revenue], SAMEPERIODLASTYEAR ( 'Date'[Date] ) ) and TOTALYTD for TODATE-style measures
What we checkThat the date table is complete and marked, and that fiscal calendars and period offsets match the Oracle time dimension.
Oracle BI patternLevel-based measureRevenue defined with its aggregation rule set at the Region level of a dimension hierarchy
Typical Power BI approachCALCULATE ( [Revenue], ALLEXCEPT ( Geography, Geography[Region] ) )
What we checkWhich levels the original measure respected, so only those filters are kept.
Oracle BI patternFILTER ... USINGFILTER("Sales"."Revenue" USING ("Products"."Brand" = 'Acme'))
Typical Power BI approachCALCULATE ( [Revenue], Product[Brand] = "Acme" )
What we checkWhether the filter should replace or add to what the user has selected.
Oracle BI patternRankingRANK("Sales"."Revenue")
Typical Power BI approachRANKX ( ALLSELECTED ( Product[Product] ), [Revenue] )
What we checkTies, sort order and the scope the ranking is computed over.
Oracle BI patternParent-child hierarchyParent-child hierarchy in the business model (for example an account or organisation tree)
Typical Power BI approachCalculated columns with PATH and PATHITEM to flatten the tree into levels, then a Power BI hierarchy
What we checkRagged branches, orphan members and how totals roll up compared with Oracle.
Oracle BI patternSession variable securityVALUEOF(NQ_SESSION."REGION") inside a data filter, populated by an initialization block
Typical Power BI approachA role filter such as Region[Region] = LOOKUPVALUE ( UserRegion[Region], UserRegion[Email], USERPRINCIPALNAME () )
What we checkThat each user sees the same rows as before, tested user by user.

Not every Oracle BI construct has a one-line equivalent. Where logic is really data preparation, such as a calculation pushed to the database, we move it into Fabric so every report reuses it.

How We Work

OBIEE to Fabric Migration: Four Phases to Cutover

A structured plan so your Oracle BI reporting moves without disruption. Oracle BI stays live while Power BI is built and validated, so users switch only when their reports are signed off.

01
Discover and Triage

We inventory your OBIEE or Oracle Analytics Server environment, check usage and owners, and agree what to retire, consolidate or rebuild, with a complexity rating that flags RPD-heavy subject areas, BI Publisher output and security.

Output: inventory and decisions
02
Model and Pilot

We build the first governed semantic model from one subject area, translate the hardest calculations and rebuild a pilot set of reports. The pilot sets DAX, naming and design standards and proves parity before the main build.

Output: signed-off pilot and standards
03
Migrate in Waves and Validate

We migrate in waves ordered by business priority and shared subject areas. Each wave is checked against the Oracle original, and the report owner signs off every report before it moves on.

Output: migrated reports with sign-off
04
Parallel Run, Cutover and Handover

We run Oracle BI and Power BI side by side, cut over by audience, retire Oracle content when you are ready and hand over documentation, runbooks and training so your team can run the platform.

Output: retired Oracle content and handover
Timeline

What the Timeline Depends On

We do not quote a duration before we have seen your estate. Length is driven by the number of reports still in scope after triage, the number and complexity of RPD subject areas, how much logic sits in calculations and variables, the number of Oracle data sources, BI Publisher volume and how long you want to run in parallel. We fix the plan in the proposal after discovery and work back from your Oracle support or renewal date where there is one.

Milestone 1
Inventory and decisions agreed

Triage complete; destination model design and wave plan approved.

Milestone 2
Pilot signed off

First semantic model live; pilot reports match Oracle BI; standards agreed.

Milestone 3
Waves complete

Each wave validated and signed off by report owners.

Milestone 4
Cutover and Oracle BI retirement

Users moved by audience; Oracle BI content retired, ideally ahead of your support or licence renewal date.

Pricing Model

How We Price an OBIEE Migration

You see real effort on your own data before committing to the full programme. Price is driven by report count, calculation complexity, subject area and data source volume, security requirements and parallel-run length.

Step 1
Free assessment

We inventory your Oracle BI estate and give you a triage, a complexity profile and a recommended scope, at no cost.

No commitment
Step 2
Fixed-scope pilot

A priced pilot that builds one semantic model and rebuilds a representative set of reports, so the approach and effort are proven on your data.

Fixed scope and fee
Step 3
Priced per wave

Remaining waves are quoted by complexity tier once the pilot gives real numbers, as fixed fee per wave or on a time and materials basis, whichever suits you.

Your choice of model
Why Numlytics

Why Choose Us for OBIEE to Fabric Migration

Six reasons teams trust us to move Oracle BI reporting to Power BI without losing the numbers their business relies on.

RPD to DAX Expertise

Translating RPD logic is where migrations stall. We map each pattern to DAX, test it at the original grain and move upstream anything that is really data preparation.

Models Before Reports

Oracle BI estates often hold several subject areas with overlapping definitions. We consolidate them into shared star-schema semantic models first, so each business metric has one definition.

Inventory Before Migration

Long-lived Oracle BI environments carry analyses nobody opens and subject areas nothing uses. We find them first, so you rebuild and test only what the business uses.

Parity You Can Audit

We do not switch reporting until totals, calculated measures and filtered views match the Oracle original and the report owner signs off.

Security Carried Across

Application roles and data filters are mapped to row-level security and tested user by user, so people see the same data in Power BI that they saw in Oracle BI.

Handover and Training

You receive architecture notes, runbooks and a team walkthrough, and we can train your Oracle BI developers in Power BI, DAX and Fabric data preparation, so your team runs the platform without us.

CASE STUDY: A TYPICAL ENGAGEMENT

The starting point: an Oracle BI environment that has grown for years. Several subject areas with overlapping definitions, an RPD that only a few people dare to change, hundreds of analyses of which few are opened, and BI Publisher reports that finance relies on. The Oracle BI version is ageing and a support or renewal date is approaching.

What we do: triage the environment, retire what nobody opens, rebuild the shared subject area as one governed semantic model, translate the calculation-heavy measures to DAX, map application roles and data filters to row-level security and rebuild the priority reports in Power BI and paginated format. Each report is checked against the Oracle original and signed off by its owner.

What the client team gets: one definition per metric, fewer reports to maintain, documented DAX, and developers trained to extend the model. Oracle BI is retired in stages once parallel running confirms the numbers hold.

FAQ

Common Questions About OBIEE to Fabric Migration

Still deciding? These are the questions teams ask most before they plan their move off Oracle BI.

Book a Free Call →
Support dates are set by Oracle per release and can change, so we do not state them here. Oracle publishes them for OBIEE in My Oracle Support Doc ID 1664916.1 and for Oracle Analytics Server in Doc ID 2775561.1. Confirm your release first. Whether to stay on Oracle, move to Oracle Analytics Server or Cloud, or move to Power BI on Microsoft Fabric is then a business decision about cost, skills and where your data platform is heading, and our free assessment sizes the Fabric option for you.
Third-party tools claim to automate parts of the conversion, and we are happy to evaluate them with you. Microsoft's migration guidance states that a conversion to Power BI cannot be completely automated from end to end. We treat automated output as a starting draft, rebuild the model on a star schema and test every report against the Oracle original.
The physical layer maps to the source tables in OneLake or your database, the business model and mapping layer becomes the star-schema semantic model with its relationships, hierarchies and DAX measures, and the presentation layer becomes display folders and perspectives. Logical SQL functions such as AGO and FILTER USING have DAX equivalents, but there is no automatic one-to-one conversion, so each is tested at the original grain.
We review each source. Where your Oracle database meets Microsoft's requirements, Fabric mirroring can replicate Oracle tables into OneLake, and Microsoft documents that it needs LogMiner, archive log mode, supplemental logging and an on-premises data gateway. Otherwise we use pipelines, copy jobs or Dataflows Gen2 with the Oracle connector. If you use Oracle BI Applications and its prebuilt warehouse, we assess that separately because it is a distinct decision.
We map each application role and data filter to row-level security. In a Power BI semantic model that means roles with filters, often driven by a user-to-region or user-to-entity table and the signed-in user, tested user by user against the Oracle behaviour. We align roles with your Microsoft Entra ID groups and choose the approach that fits your model and storage mode.
Pixel-perfect BI Publisher output is rebuilt as Power BI paginated reports, and templates do not convert automatically, so we rebuild each one that is still needed. Agents and scheduled delivery are replaced with Power BI subscriptions and alerts per report, which we plan during triage.
Timing and cost depend on the number of reports in scope after triage, the number and complexity of subject areas, how much logic sits in the RPD and the volume of BI Publisher output. We start with a free assessment, then a fixed-scope pilot, and quote the remaining waves from real pilot numbers. Oracle BI stays live during the build and parallel run, users move by audience once their reports are signed off, and we provide handover documentation and optional corporate Power BI training.
Get Started

Start Your OBIEE to Fabric Migration Plan

Book a free assessment. We will inventory your RPD, analyses, dashboards and BI Publisher reports, flag the calculation-heavy content and recommend what to rebuild, consolidate or retire, at no cost.