CSV Validation Example

CsvValidationExample.cs

// C# 11 and .NET 6+

using Sylvan.Data;     // 0.2.12-B0001
using Sylvan.Data.Csv; // 1.2.7

// the schema for the csv data below.
var schema =
    new Schema.Builder()
    // ID is required!
    .Add<int>("Id")
    .Add<string>("Name", allowNull: false)
    // Date of birth can be null.
    .Add<DateTime?>("DOB")
    // nullable enum value
    .Add<ConsoleColor?>("FavoriteColor")
    .Build();

// record 1 is OK
// record 2 has a missing ID and an unknown color. The different date format is OK.
// record 3 has a bad date (Unknown). The missing/null color is OK, as it is nullable.
var data =
    """
    Id,Name,DOB,FavoriteColor
    1,Dan,2020-01-01,Red
    ,Alex,"Jun 29, 1995",Blornge
    3,Maria,Unknown
    """;

// create a reader with the schema
var opts = new CsvDataReaderOptions { Schema = new CsvSchema(schema) };
var reader = CsvDataReader.Create(new StringReader(data), opts);

// apply the validation handler. See "HandleRecordError" function below
var validatingReader = reader.ValidateSchema(HandleRecordError);

// This will produce records for row 1, which had no errors
// and row 3 where the invalid date could be corrected.
foreach (var record in validatingReader.GetRecords<Record>())
{
    Console.WriteLine(record.ToString());
}

// the validation handler will log all failures
// it will repair bad Date values by setting the value to null
// Any records that can be repaired will be returned to the reader.
static bool HandleRecordError(SchemaValidationContext context)
{
    var csv = (CsvDataReader)context.DataReader;

    Console.WriteLine("--- ERROR START ---");
    Console.WriteLine($"Error(s) on row {csv.RowNumber}. Raw Record:");
    Console.Out.Write(csv.GetRawRecordSpan());
    Console.WriteLine("");

    bool repaired = true;
    foreach (var idx in context.GetErrors())
    {
        var ex = context.GetException(idx);
        // log the error detail
        Console.WriteLine($"   col: {idx} value: \"{csv.GetString(idx)}\" exception:  {ex.Message} ({ex.GetType()})");

        var name = csv.GetName(idx);
        switch (name)
        {
            case "DOB": // the date column
                // unknown date, we'll set it to null
                context.SetValue(idx, null);
                break;
            default:
                // any other issue is irreparable
                repaired = false;
                break;
        }
    }

    Console.WriteLine(repaired ? "Record was repaired" : "Record was NOT repaired");
    Console.WriteLine("--- ERROR END ---");

    return repaired;
}

class Record
{
    public int Id { get; set; }
    public string Name { get; set; }
    public DateTime? DOB { get; set; }
    public ConsoleColor? FavoriteColor { get; set; }

    public override string ToString()
    {
        return $"Record: {Id} {Name} {DOB?.ToString() ?? "NULL"} {FavoriteColor}";
    }
}