Reading a CSV file without strongly typed structure

[viksing]

How can I read a csv to data table, and then use it in something like

var dataView = mlContext.Data.LoadFromEnumerable(new List { dynamicDataObject })

Why Do I need to provide a concrete class for the type of the data?

I tired to use DynamicData like var dynamicDataObject = new DynamicData { DataTable = dynamicDataTable };

but it doesn't work. I do not know how to create SchemaDefinition.

static SchemaDefinition GenerateSchemaDefinition(Type dataType)
{
    var properties = dataType.GetProperties(BindingFlags.Public | BindingFlags.Instance);

    var columns = properties.Select(p => new SchemaDefinition.Column
    {
        ColumnName = p.Name,
        ColumnType = p.PropertyType
    }).ToArray();

    return new SchemaDefinition(columns);
}

this says p.PropertyType is not of type DataViewType.

It is  just too complex.

---

Document Details

⚠ Do not edit this section. It is required for learn.microsoft.com ➟ GitHub issue linking.

• ID: 10e8d5aa-2038-a382-ec86-cf8ad031f996
• Version Independent ID: 81198dc9-ddf4-cef1-c178-82f50f0e7a46
• Content: DataOperationsCatalog.LoadFromEnumerable Method (Microsoft.ML)
• Content Source: dotnet/xml/Microsoft.ML/DataOperationsCatalog.xml
• Product: dotnet-ml-api
• GitHub Login: @natke
• Microsoft Alias: nakersha

[asmirnov82]

Hi @viksing, have you tried to use a dataframe for this?

IDataView dv = DataFrame.LoadCsv("filename.csv");

You may find more info about using DataFrame with ML .Net here: https://learn.microsoft.com/en-us/dotnet/machine-learning/how-to-guides/getting-started-dataframe