Yres and Microsoft Fabric: competitor or combination?
Microsoft Fabric and Yres do not compete; they sit on different layers. Fabric is a platform: storage, compute and a growing set of building blocks. Yres is an automation tool that builds and maintains the data warehouse: loading sources, keeping history, deploying changes, monitoring. Today Yres runs on Azure SQL and delivers a Parquet feed that Fabric can read directly. Fabric as a target platform for Yres is planned for the end of 2027.
Last updated:
What Fabric does and does not do for you
Fabric bundles storage (OneLake), compute and tooling: pipelines, dataflows, a warehouse, a lakehouse and Power BI, in one environment and one licence. That makes it attractive for organisations already on Microsoft. It is not, however, a data warehouse that builds itself: which sources you load, how you keep history, how changes reach production and who checks at night whether everything went well, you still assemble yourself.
The building blocks do keep getting richer. This is what Microsoft itself documents, as of 21 September 2026:
| Component | What Fabric brings | What you still do yourself |
|---|---|---|
| Incremental loading | Copy job can copy incrementally on a watermark or through change data capture | Configure per source, choose a watermark per table, handle failures |
| History (SCD2) | Copy job has a built-in SCD2 write method, in preview, CDC sources only and net changes only; otherwise a do-it-yourself pattern in Dataflow Gen2 | Build history for non-CDC sources (most SaaS APIs) yourself |
| Mirroring Azure SQL to OneLake | Mirroring, generally available, replication is free; delivers a read-only current copy | Not possible below 100 DTU (Free, Basic, Standard S0–S2), not with CDC enabled; no history |
| Connectors | Generic REST, OData, HTTP, SAP Table, SAP BW, SAP HANA | No AFAS connector; Exact Online only through Exact's own Premium connector in Dataflow Gen2, with an on-premises gateway |
| Change management | Deployment pipelines and Git integration per workspace | What a change consists of, dependencies, and the audit of what someone changes directly |
| Monitoring | Monitoring hub per run | Health checks on the configuration, a test suite, retention of your own logs |
How Yres connects to Fabric today
Yres loads your sources into an Azure SQL database in your own tenant and keeps the full history there. In addition, if you enable it, Yres writes a feed of every change as Parquet files to your Data Lake: per table, per load, marked as new, changed or deleted. Those files can be read without detour in a Fabric Warehouse or Lakehouse; for Direct Lake in Power BI you fold them into a Delta table with a notebook.
In practice that means: loading, history, the AFAS and Exact connections, change management and monitoring you leave to Yres, and you use Fabric for what it is good at: compute, notebooks, data science and Power BI on OneLake. The source of truth stays the Azure SQL database; the feed is a delivery layer.
- Fabric Warehouse: read the Parquet folder as an external table, no Spark needed
- Fabric Lakehouse and notebooks: merge the feed into a Delta table, for example for Direct Lake
- Mirroring the Yres database to OneLake also works, on an Azure SQL tier of at least 100 DTU and as long as CDC is off; you then get the current state, the history sits in the tables themselves
On the roadmap: Fabric as a target platform
By the end of 2027 Yres is planned to be able to place the data warehouse directly in Fabric, with the SQL database in Fabric as the target instead of Azure SQL. Loading, history and change management stay as they are; storage moves to OneLake, where Power BI can sit on top through Direct Lake. Until then the combination above is the recommended route. Organisations that start with Yres now do not have to rebuild later: the model and the configuration carry over.
When Fabric alone, and when Yres alongside it
Fabric alone suits an organisation with a data platform team of its own that wants to build and maintain the loading, the history and the operations itself, and with sources Fabric has good connectors for. Yres alongside it suits you as soon as AFAS, Exact Online or SAP is a main source, the team is small, or you want loading, history, change management and monitoring taken care of as standard rather than built. Fabric's licence cost does not change either way; what changes is the number of hours you spend building and operating.
Frequently asked questions
Does Fabric replace a tool like Yres?
No. Fabric provides the platform; Yres builds and maintains the data warehouse on or next to it. Fabric's building blocks keep getting richer, but assembling them into a reliable whole remains work you do yourself or leave to a tool.
Can I use the Yres data in Fabric today?
Yes, in two ways: through the Parquet feed to your Data Lake, which a Fabric Warehouse or Lakehouse reads directly, or by mirroring the Yres Azure SQL database to OneLake. Mirroring requires an Azure SQL tier of at least 100 DTU and does not work with change data capture enabled.
Does Fabric have an AFAS or Exact connector?
No AFAS connector, per Microsoft's connector overview as of 21 September 2026. For Exact Online there is a connector supplied by Exact itself, in Dataflow Gen2 only, requiring an Exact Online Premium subscription and an on-premises gateway. Yres has its own connection for both and is an official partner of both.
When can Yres place the data warehouse in Fabric itself?
That is planned for the end of 2027. Until then Yres runs on Azure SQL and connects to Fabric through the Parquet feed or mirroring. Whoever starts now does not rebuild later: the model and the configuration carry over.
