Migration
Informatica PowerCenter to Microsoft Fabric migration: a practical guide
How to move Informatica PowerCenter mappings and workflows to Microsoft Fabric: choosing Dataflow Gen2, Warehouse SQL or Spark notebooks per mapping, transformation mapping, pipelines for workflows, and reconciliation.
Hemant Walchale
· 2 min read
For Microsoft-standardized organizations, Fabric is a natural home for the transformation logic that Informatica PowerCenter runs today: it removes a separate ETL licence and puts pipelines, data and Power BI in one platform. Fabric offers three ways to express transformation logic, and choosing the right one per mapping is the key decision.
This is the approach we recommend.
1. Export and inventory
Export the repository as XML with pmrep or Repository Manager, count transformation types across the estate, and use workflow run history to find what still runs. Note Java, SQL and Stored Procedure transformations, Update Strategy logic, mapplets, parameter files and pre- and post-session SQL. Retire unused mappings first.
2. Pick a Fabric engine per mapping
| Mapping profile | Fabric engine |
|---|---|
| Simple, low volume, maintained by analysts | Dataflow Gen2 (Power Query) |
| Set-based logic over large tables, SQL-skilled team | T-SQL stored procedures in a Fabric Warehouse |
| Complex logic, very large data, files or semi-structured data | Spark notebooks in a Lakehouse |
3. Map transformations
| PowerCenter | In Fabric |
|---|---|
| Source Qualifier | Copy activity or mirroring into OneLake; SQL overrides become the first query |
| Expression, Filter, Router | SQL expressions and WHERE, Dataflow columns and filters, or DataFrame logic |
| Joiner, Lookup | Joins, or merge queries in Dataflow Gen2 |
| Aggregator, Sorter, Rank | GROUP BY and window functions |
| Update Strategy | MERGE in the Warehouse where supported, or Delta merge in Spark |
| Sequence Generator | Identity columns where supported, or keys generated in the pipeline |
| Mapplets | Shared SQL views, procedures or notebook functions |
| Java, SQL, Stored Procedure transformations | Rewrite in T-SQL or Python; test each |
4. Rebuild workflows as pipelines
Sessions become activities, links become dependencies (success, failure, completion), Decision tasks become If Condition activities, worklets become child pipelines, Event-Wait tasks become event triggers, Email tasks become Outlook or Teams activities and the scheduler becomes pipeline schedules.
5. Reconcile and retire
Run PowerCenter and Fabric side by side and compare every target table. Watch implicit type conversions, null handling in lookups, trailing spaces, decimal precision and session-level date defaults, plus Fabric Warehouse’s case-sensitive comparisons. Cut over by domain and plan back from the licence renewal.
The short version
Export and count transformation types, choose Dataflow Gen2, Warehouse SQL or Spark per mapping, map transformations, rebuild workflows as Data Factory pipelines, reconcile every table and switch off PowerCenter domain by domain.
Compare with Informatica to Databricks and Informatica to Snowflake, or browse all migration guides.
Not sure where your own platform stands? Take the free 3-minute data platform scorecard. Unfamiliar terms are explained in the data engineering glossary.