dotnet/machinelearning · critical · ArgumentNullException

Provided path cannot be null

Error message

Provided path cannot be null

What it means

Thrown by UserInputValidationUtil.ValidatePath when the file path supplied to the infer-columns API (ValidateInferColumnsArgs) is null. The path identifies the training data file to inspect, so it is mandatory and reported via ArgumentNullException naming the path parameter.

Source

Thrown at src/Microsoft.ML.AutoML/Utils/UserInputValidationUtil.cs:168

            if (columns?.Contains(null) == true)
            {
                throw new ArgumentException($"Null column string was specified as {columnPurpose} in column information");
            }
        }

        private static void ValidateLabelColumn(string labelColumn)
        {
            if (labelColumn == null)
            {
                throw new ArgumentException("Provided label column cannot be null");
            }
        }

        private static void ValidatePath(string path)
        {
            if (path == null)
            {
                throw new ArgumentNullException(nameof(path), "Provided path cannot be null");
            }

            var fileInfo = new FileInfo(path);

            if (!fileInfo.Exists)
            {
                throw new ArgumentException($"File '{path}' does not exist", nameof(path));
            }

            if (fileInfo.Length == 0)
            {
                throw new ArgumentException($"File at path '{path}' cannot be empty", nameof(path));
            }
        }

        private static void ValidateValidationData(IDataView trainData, IDataView validationData)
        {
            if (validationData == null)

View on GitHub (pinned to 7b76e69cf9)

Solutions

1. Provide a valid file path string to the InferColumns call
2. Resolve the path from configuration with a null/empty check before calling
3. Fix the config/CLI plumbing so the path value is populated

Example fix

// before
string dataPath = configuration["DataPath"]; // null
var columns = mlContext.Auto().InferColumns<ModelInput>(dataPath, labelColumnName: "Label");

// after
string dataPath = configuration["DataPath"] ?? throw new InvalidOperationException("DataPath config missing");
var columns = mlContext.Auto().InferColumns<ModelInput>(dataPath, labelColumnName: "Label");

if (string.IsNullOrWhiteSpace(path)) throw new InvalidOperationException("Data file path is required");

static bool HasPath(string p) => !string.IsNullOrWhiteSpace(p);

try { var cols = mlContext.Auto().InferColumns(path, labelColumnName); }
catch (ArgumentNullException ex) when (ex.ParamName == "path") { /* resolve path from fallback config and retry */ }

When it happens

Trigger: Calling InferColumns overloads with path = null, usually because a config value or command-line argument was not supplied.

Common situations: Missing appsetting/connection-string entry for the data file path; CLI args array missing an element; environment-specific path variable empty in CI.

Related errors

• Training data cannot be null
• Provided label column cannot be null
• File '{path}' does not exist
• If provided, {nameof(samplingKeyColumnName)} must be the sam
• Training data has 0 rows

AI-assisted analysis of dotnet/machinelearning@7b76e69cf9 (2026-09-11). Data as JSON: /api/errors/e40c02ee57c2bddb. Report an issue: GitHub.