NullReferenceException when trying to run AutoML regression experiment from code

[superichmann]

System Information (please complete the following information):

• OS & Version: Microsoft Windows 10 Pro 10.0.19045 N/A Build 19045
• ML.NET Version: 2.0.1
• microsoft.ml.automl: 0.20.1
• .NET Version: .NET 6.0

Describe the bug
When trying to run AutoML regression experiment on a table with data of 30 rows in QuestDB (implements Postgresql) the following exception occurs:

	Message	"`Object reference not set to an instance of an object.`"	string
	Source	"`Microsoft.ML.AutoML`"	string
	StackTrace	" ` at Microsoft.ML.AutoML.AutoMLExperiment.<RunAsync>d__26.MoveNext()`"	string
	TargetSite	`{Void MoveNext()}	System.Reflection.MethodBase {System.Reflection.RuntimeMethodInfo}`
	DeclaringType	{Name = "`<RunAsync>d__26" FullName = "Microsoft.ML.AutoML.AutoMLExperiment+<RunAsync>d__26`"}	System.Type {System.RuntimeType}

To Reproduce
Download QuestDB
Create table "CREATE TABLE 'tests' (Ticker SYMBOL, H FLOAT, P_T FLOAT,timestamp TIMESTAMP) timestamp (timestamp) PARTITION BY MONTH;"
Insert 30 rows to the table
IN C#: get the table data into IDataView with mlContext.Data.CreateDatabaseLoader and using Npgsql;
run the following code:
ExperimentResult<RegressionMetrics> experimentResult = mlContext.Auto().CreateRegressionExperiment(20).Execute(dataview, "P_T");
Expected behavior
The code should not throw an Exception and finish the experiment. This code works perfectly on a large table with 28000 rows and 30 columns

Additional context
I Tried running the experiment from the modelbuilder on csv and it works.

[superichmann]

I have tested same data csv load on C# code and got the same exception :(

[superichmann]

questdb-query-1675793575444.csv

[LittleLittleCloud]

This looks like a dup of #6446

The fix has been available on nightly build. It still needs some time to go into nuget though

[superichmann]

mmm if you want I cen test it just tell me how

[LittleLittleCloud]

Sure. The package is available on nightly build. You can find the nightly build feed on readme and simply install the latest version

[superichmann]

I tried running the instructions here for half a day now and I got much errors in the type of The source code won't be available via Source Link. :((((((((

Maybe for now you can tell me an approximation of minimum data size that will not trigger this exception?

[LittleLittleCloud]

Looks like you are building from source? You don't have to do it if you just want to use the nightly build package.

Here's the instruction of how to add nightly build feed to your project. ( You should not need to install Azure Artifacts credential provider for the nightly build feed since it is public and doesn't require credential)
https://dev.azure.com/dnceng/public/_artifacts/feed/dotnet-libraries/connect/dotnet

[superichmann]

I have done these instructions (Except for restore packages) and installed from NuGet manager from dotnet-libraries source, still getting these versions :(

microsoft.ml\3.0.0-preview.23109.1
microsoft.ml.automl\0.21.0-preview.23109.1

[LittleLittleCloud]

@superichmann

Looks like you are on the right nightly build package. Does your code work now.

[superichmann]

No, still same exception.

As well I have tested on these versions since they are in the prealpha tick box on the regular NuGet feed... but I thought these are not the ones you meant

[superichmann]

System.AggregateException
 HResult=0x80131500
 Message=One or more errors occurred. (Object reference not set to an instance of an object.)
 Source=System.Private.CoreLib
 StackTrace:
 at System.Threading.Tasks.Task.ThrowIfExceptional(Boolean includeTaskCanceledExceptions)
 at System.Threading.Tasks.Task`1.GetResultCore(Boolean waitCompletionNotification)
 at Microsoft.ML.AutoML.AutoMLExperiment.Run()
 at Microsoft.ML.AutoML.RegressionExperiment.Execute(IDataView trainData, ColumnInformation columnInformation, IEstimator`1 preFeaturizer, IProgress`1 progressHandler)
 at AutoML_Example.Program.Main(String[] args) in source\repos\ModelBuilder001\Program.cs:line 29

 This exception was originally thrown at this call stack:
 [External Code]

Inner Exception 1:
NullReferenceException: Object reference not set to an instance of an object.

[LittleLittleCloud]

Hmmm interesting, I test your dataset on main branch and that works

Here's my code

public void AutoFit_QuestDB()
 {
 var context = new MLContext(1);
 context.Log += (o, e) =>
 {
 if (e.Source.StartsWith("AutoMLExperiment"))
 {
 this.Output.WriteLine(e.RawMessage);
 }
 };
 var data = context.Data.LoadFromTextFile<QuestDB>("C:\\Users\\xiaoyuz\\Desktop\\questdb-query-1675793575444.csv", hasHeader: true, separatorChar: ',');
 var label = "P_T";
 var settings = new RegressionExperimentSettings
 {
 MaxModels = 100,
 };


 var result = context.Auto()
 .CreateRegressionExperiment(settings)
 .Execute(data, 10, label);

 Assert.True(result.BestRun.Results.Select(x => x.ValidationMetrics.RSquared).Min() > 0.70);
 }

class QuestDB
 {
 [LoadColumn(0)]
 [ColumnName("Ticker")]
 public string Ticker { get; set; }

 [ColumnName("H")]
 [LoadColumn(1)]
 public float H { get; set; }

 [ColumnName("P_T")]
 [LoadColumn(2)]
 public float PT { get; set; }

 [ColumnName("timestamp")]
 [LoadColumn(3)]
 public DateTime TimeStamp { get; set; }
 }

[LittleLittleCloud]

Can you give me a quick look of your csproj file

[superichmann]

omg funny! here it is:

<Project Sdk="Microsoft.NET.Sdk">
 <PropertyGroup>
 <OutputType>Exe</OutputType>
 <TargetFramework>net6.0</TargetFramework>
 <ImplicitUsings>enable</ImplicitUsings>
 <Nullable>enable</Nullable>
 </PropertyGroup>
 <ItemGroup>
 <PackageReference Include="Microsoft.ML" Version="3.0.0-preview.23109.1" />
 <PackageReference Include="Microsoft.ML.AutoML" Version="0.21.0-preview.23109.1" />
 </ItemGroup>
</Project>