-
Notifications
You must be signed in to change notification settings - Fork 1.9k
Description
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.