D365FSCM Demand forecasting using Azure ML service output parsing error

Shubhadeep Bhowmick 0 Reputation points


I have completed Demand forecasting setup in D365 and the pipeline is executing successfully and output files are getting stored in demplan-azureml/outputs folder.
But in Generate forecast in Azure Machine Learning service step it is showing error "Demand forecast cannot be generated in Azure Machine Learning service. The following error has occurred: The value could not be parsed."
From the detailed log it looks like the error is coming from D365 side.

Sample output file using demo data -
"usmf 1 D_Aloc D0004 000005 ",1459,148,16.6308458921961,-1,ARIMA

"usmf 1 D_Aloc D0004 000005 ",1460,148,16.6308458921961,0,

"usmf 1 D_Aloc D0004 000005 ",1461,148,16.6308458921961,0,

"usmf 1 D_Aloc D0004 000005 ",1462,148,16.6308458921961,0

detail log :

System.Reflection.TargetInvocationException: Exception has been thrown by the target of an invocation. ---> System.FormatException: The value could not be parsed.
   at System.Numerics.BigNumber.ParseBigInteger(String value, NumberStyles style, NumberFormatInfo info)
   at Microsoft.Dynamics.AX.DemandPlanning.Azure.Data.BaseReader.GetDecimal(Int32 i)
   at Dynamics.AX.Application.ReqDemPlanAzureMlStrategy.prepareForecastLine(ReqDemPlanForecast _forecast, ReqDemPlanCreateForecastDataContract _dataContract, ISimpleDataReader _reader, ReqDemPlanDateConverter _dateConverter, List _enabledAttributes, ReqDemPlanAllocationKeyFilterTmp _allocationKeyFilterTable)
   at Dynamics.AX.Application.ReqDemPlanAzureMlStrategy.importPredictions(ReqDemPlanCreateForecastDataContract _dataContract, ReqDemPlanAllocationKeyFilterTmp _allocationKeyFilterTable)
   at Dynamics.AX.Application.ReqDemPlanAzureMlStrategy.generateAndImportPredictions(ReqDemPlanCreateForecastDataContract _dataContract, ReqDemPlanAllocationKeyFilterTmp _allocationKeyFilterTable)
   --- End of inner exception stack trace ---
Azure Machine Learning
Azure Machine Learning
An Azure machine learning service for building and deploying models.
2,615 questions
{count} votes