Passing in expected input_data to endpoint returns ForecastingDataException

Question

I created an online endpoint for a model trained by AutoML (so the score python file was internally built by AutoML). When I pass a JSON input to the endpoint with the same parameters as the column names in the training CSV file, I get this error. What am I doing wrong?

azureml.automl.core.shared.forecasting_exception.ForecastingDataException: ForecastingDataException:
	Message: Argument [data] is of unsupported type: []. Supported type(s): [pandas.DataFrame]
	InnerException: None
	ErrorResponse 
{
    "error": {
        "code": "UserError",
        "message": "Argument [data] is of unsupported type: []. Supported type(s): [pandas.DataFrame]",
        "details_uri": "https://aka.ms/AutoMLConfig",
        "target": "data",
        "inner_error": {
            "code": "BadArgument",
            "inner_error": {
                "code": "ArgumentInvalid"
            }
        },
        "reference_code": "ac316505-87e4-4877-a855-7532473a796c"
    }
}

{"input_data":{"Date":"2024-04-02", "P1":935.5, "P2":65, "P3":45, "P4":38, "P5":51, "P6":17, "P7":38}}

Accepted Answer

Hi @Sunand Sandurkar

Thank you for using the Microsoft Q&A forum.

Based on the error message, it seems that you are trying to pass a dictionary object as input to your endpoint, which is not supported as the endpoint seems to expect a pandas DataFrame object as input and not a dictionary object, by its sheer design.

For your use case you convert the input dictionary to a pandas DataFrame object and then Pass the DataFrame object as input to the endpoint.

For more details, please refer.

I hope you understand. Thank you.

Please do not forget to click Accept Answer and Yes for was this answer helpful, wherever the information provided helps you. This can be beneficial to other community members.

Share via

Passing in expected input_data to endpoint returns ForecastingDataException

0 additional answers

Your answer