Forum Discussion
Exporting data from Semantic model
- 1 year ago
Hi SalvaC ,
You're in a tricky but not uncommon situation — pulling data from an OData source (Jira) into Power BI works fine with token-based auth, but exporting it out of the semantic model (or transforming it via SQL) is tough in an on-premises-only setup without Fabric or cloud-based tools like Power Automate. Here's the core issue:
-
Power BI semantic models are not designed for downstream ETL consumption directly.
-
You can't export or connect SSIS to semantic models without going through Power BI Service or Fabric, which your client can't use.
-
SSIS doesn’t support blank usernames for token-based OData access natively — that’s the blocker with going direct from Jira to SSIS.
Practical Options:
-
Use Power BI Desktop as a "bridge":
-
Load Jira OData into Power BI Desktop (as you already do).
-
Export data from the Power BI model to flat files (CSV) using DAX Studio or Tabular Editor.
-
Then, have SSIS pick up those flat files for processing.
-
This can be automated with scripting and scheduled tasks.
-
-
Use PowerShell to Extract from the Semantic Model:
-
With Power BI Desktop running, connect to the local model via localhost:xxxxx and use PowerShell with TOM (Tabular Object Model) or DAX Studio to extract data.
-
Save to flat files or push to your SQL database from PowerShell.
-
-
Use a Middleware ETL Tool with OData Support:
-
Tools like KingswaySoft (SSIS add-on) allow advanced authentication scenarios for OData, including custom headers (so you can manually set the token as the Authorization: Basic header).
-
This bypasses the username restriction in native SSIS.
-
-
Use Custom .NET/PowerShell App to Pull and Push:
-
Build a lightweight custom app or script that pulls data from Jira (via the OData feed with token auth) and pushes it into SQL Server — effectively replacing SSIS for this job.
-
Once in SQL Server, you can continue the rest of your ETL via SSIS.
-
Not Viable (without Fabric):
-
Dataflows Gen2 – require Fabric capacity.
-
Power Automate export from semantic model – also needs Fabric.
-
Direct SSIS to Power BI semantic model or dataset – not supported.
-
Hi SalvaC ,
You're in a tricky but not uncommon situation — pulling data from an OData source (Jira) into Power BI works fine with token-based auth, but exporting it out of the semantic model (or transforming it via SQL) is tough in an on-premises-only setup without Fabric or cloud-based tools like Power Automate. Here's the core issue:
-
Power BI semantic models are not designed for downstream ETL consumption directly.
-
You can't export or connect SSIS to semantic models without going through Power BI Service or Fabric, which your client can't use.
-
SSIS doesn’t support blank usernames for token-based OData access natively — that’s the blocker with going direct from Jira to SSIS.
Practical Options:
-
Use Power BI Desktop as a "bridge":
-
Load Jira OData into Power BI Desktop (as you already do).
-
Export data from the Power BI model to flat files (CSV) using DAX Studio or Tabular Editor.
-
Then, have SSIS pick up those flat files for processing.
-
This can be automated with scripting and scheduled tasks.
-
-
Use PowerShell to Extract from the Semantic Model:
-
With Power BI Desktop running, connect to the local model via localhost:xxxxx and use PowerShell with TOM (Tabular Object Model) or DAX Studio to extract data.
-
Save to flat files or push to your SQL database from PowerShell.
-
-
Use a Middleware ETL Tool with OData Support:
-
Tools like KingswaySoft (SSIS add-on) allow advanced authentication scenarios for OData, including custom headers (so you can manually set the token as the Authorization: Basic header).
-
This bypasses the username restriction in native SSIS.
-
-
Use Custom .NET/PowerShell App to Pull and Push:
-
Build a lightweight custom app or script that pulls data from Jira (via the OData feed with token auth) and pushes it into SQL Server — effectively replacing SSIS for this job.
-
Once in SQL Server, you can continue the rest of your ETL via SSIS.
-
Not Viable (without Fabric):
-
Dataflows Gen2 – require Fabric capacity.
-
Power Automate export from semantic model – also needs Fabric.
-
Direct SSIS to Power BI semantic model or dataset – not supported.
Hi all,
Thank you very much for all the suggestions.
You are awesome.
We have looked into it and will attempt to retrieve the data directly from the Jira database, which is stored on a SQL Server Instance.
It's not my preferred solution, as it requires testing whenever the software is updated.
Even some add-ons will potentially change the database.
But the restriction on place leaves us with little choice.
Thank you again.
Kind regards
Salvatore