dotnet/machinelearning · error · ArgumentException

Training data has 0 rows

Error message

Training data has 0 rows

What it means

Thrown by UserInputValidationUtil.ValidateTrainData when the supplied IDataView is non-null but contains zero rows, as determined by DatasetDimensionsUtil.IsDataViewEmpty. AutoML cannot train or split an empty dataset, so the input is rejected with an ArgumentException naming trainData.

Source

Thrown at src/Microsoft.ML.AutoML/Utils/UserInputValidationUtil.cs:78

        public static void ValidateSamplingKey(string samplingKeyColumnName, string groupIdColumnName, TaskKind task)
        {
            if (task == TaskKind.Ranking && samplingKeyColumnName != null && samplingKeyColumnName != groupIdColumnName)
            {
                throw new ArgumentException($"If provided, {nameof(samplingKeyColumnName)} must be the same as {nameof(groupIdColumnName)} for Ranking Experiments", samplingKeyColumnName);
            }
        }

        private static void ValidateTrainData(IDataView trainData, ColumnInformation columnInformation)
        {
            if (trainData == null)
            {
                throw new ArgumentNullException(nameof(trainData), "Training data cannot be null");
            }

            if (DatasetDimensionsUtil.IsDataViewEmpty(trainData))
            {
                throw new ArgumentException("Training data has 0 rows", nameof(trainData));
            }

            foreach (var column in trainData.Schema)
            {
                if (column.Name == DefaultColumnNames.Features && column.Type.GetItemType() != NumberDataViewType.Single)
                {
                    throw new ArgumentException($"{DefaultColumnNames.Features} column must be of data type {NumberDataViewType.Single}", nameof(trainData));
                }

                if ((column.Name != columnInformation.LabelColumnName &&
                    column.Name != columnInformation.UserIdColumnName &&
                    column.Name != columnInformation.ItemIdColumnName &&
                    column.Name != columnInformation.GroupIdColumnName)
                    &&
                        column.Type.GetItemType() != BooleanDataViewType.Instance &&
                        column.Type.GetItemType() != NumberDataViewType.Single &&
                        column.Type.GetItemType() != TextDataViewType.Instance)
                {

View on GitHub (pinned to 7b76e69cf9)

Solutions

1. Check trainData.GetRow count / row cursor before executing; ensure the dataset has at least one row
2. Fix the filter/query that eliminated all rows
3. Verify the source file or enumerable actually contains data rows, not just a header

Example fix

// before
var data = mlContext.Data.LoadFromEnumerable(items.Where(x => x.Year == requestedYear));
var result = experiment.Execute(data, nameof(ModelInput.Label), "A");

// after
var filtered = items.Where(x => x.Year == requestedYear).ToList();
if (filtered.Count == 0) throw new InvalidOperationException("No training rows match the filter");
var data = mlContext.Data.LoadFromEnumerable(filtered);
var result = experiment.Execute(data, nameof(ModelInput.Label), "A");

long rowCount = 0;
using (var cur = trainData.GetRowCursor(trainData.Schema))
    while (cur.MoveNext()) { rowCount++; break; }
if (rowCount == 0) throw new InvalidOperationException("Training data has no rows");

try { var r = experiment.Execute(data, label, "A"); }
catch (ArgumentException ex) when (ex.Message.Contains("0 rows")) { /* surface upstream ETL failure */ }

When it happens

Trigger: Passing an IDataView built from an empty file, an empty DataFrame, a filtered view (e.g. after a FilterRowsByColumn predicate excluding everything), or an empty enumerable via LoadFromEnumerable.

Common situations: CSV with only a header row; date-range or category filters that exclude all rows at runtime; upstream ETL job produced no output; test fixture files empty after truncation.

Understand the failure class

Background: "must not be empty", "cannot be empty" — required-field validation errors across open-source libraries — this error's family across 41 libraries.

Related errors

• Training data cannot be null
• If provided, {nameof(samplingKeyColumnName)} must be the sam
• {DefaultColumnNames.Features} column must be of data type {N
• Only supported feature column types are {BooleanDataViewType
• Duplicate column name {duplicateColName} is present in two o

AI-assisted analysis of dotnet/machinelearning@7b76e69cf9 (2026-09-11). Data as JSON: /api/errors/04ad29e6fc0d1e30. Report an issue: GitHub.