Power BI Output Tool (Alteryx) - Bad request syntax or unsupported method error when DateT

[Junato47]

We are using the Power BI Output Tool in Alteryx Designer to publish data to the Power BI Service through the REST API.

When the workflow contains a DateTime field, the upload fails with the following error:
Bad request syntax or unsupported method. Error parsing request for dataset. Unexpected character encountered while parsing value.

The same workflow publishes successfully when the DateTime field is removed.

Troubleshooting and Tests Performed:

Converted DateTime to String

Tried field types: V_String, V_WString

Formats tested: MM/dd/yyyy, yyyy-MM-dd HH:mm:ss, yyyy-MM-ddTHH:mm:ssZ, yyyy-MM-ddTHH:mm:ss.000Z

All resulted in the same error

Tested Locale Settings

Changed both Alteryx and Power BI locales to en-US and en-GB

No change in behavior

Created a New Dataset

Published to a new dataset name to avoid schema mismatch

Same error occurred

Split Date and Time into Separate Fields

Sent Date and Time separately as strings

Still failed

Removed milliseconds and timezone tokens

Used format yyyy-MM-ddTHH:mm:ss without .000Z

Failed again

Removed DateTime field completely

Workflow succeeded immediately

Environment:

Alteryx Designer version: [insert version]

Power BI Output Tool version: [insert version]

Power BI Service workspace type: [Pro / Premium / Fabric]

Request:
Please confirm if there is a known issue or restriction with the Power BI REST API (used by the Power BI Output Tool) when handling DateTime or String-formatted date fields.
We need clarification on the accepted DateTime or text format for push datasets.

[alexnajm · Accepted solution]

Yes there is a restriction documented here: Microsoft Power BI Output Tool. "Due to Power BI API limitation, Designer can't write DateTime values. Convert the DateTime values to strings before uploading the dataset to Power BI."

[Junato47]

hanks for confirming and for sharing the documentation link. Yes, we were able to successfully push the DateTime field as a string and read it in Power BI.

However, the challenge now is that we’re not able to change the data type of that field inside Power BI, since the dataset is connected through a Live Connection created by the Power BI Output Tool.

I think I’ll raise this separately in the Power BI Community to understand why this behavior exists — specifically, whether there’s any way to:

Convert this dataset connection to Import mode or Direct Query, or

Change the field type to Date/DateTime within Power BI after publishing.

If you happen to know of any workaround or configuration that allows that, please let me know — it would be very helpful.

Thanks again for your quick response and for pointing me to the documentation.

[Junato47]

Thanks for the clarification earlier.

I also checked this behavior with the Power BI team, and this is the response I received from the Power BI Community:

“This behavior is expected with the Alteryx to Power BI REST API route.
When you publish through the Power BI Output Tool, the dataset is created as a push dataset (streaming/live connection). Push datasets do not have a Power Query layer, so you cannot transform columns or change data types inside Power BI. That is why the Transform Data option is disabled and why the dataset cannot be switched to Import or DirectQuery mode afterward.
The recommended approach is to push the data into a Power BI Dataflow, then build the report on top of that. Dataflows support transformations and proper DateTime types.”

Based on this, I wanted to check whether the Power BI Output Tool in Alteryx can publish directly into a Power BI Dataflow instead of a push dataset.

Is it possible for the Output Tool to connect to a Dataflow or write into one?
If yes, we can test that path to see if it helps handle the DateTime column correctly.

Thanks again for the support so far — any guidance on whether the Dataflow route is supported would be very helpful.

[alexnajm]

I don't think this option would be supported, so I would look at outputting to a location that your Power BI dashboards can reference - hopefully then you can transform as needed.