Ignore columns from a dynamic records list

[klaudyuxxx]

I'm using Dapper to Query an SQL procedure to which I don't know the returned columns.
I'm trying to configure CSVHelper to ignore some of the columns returned by the SQL Procedure, before writing them to a file.
Can you please point me to the right direction?

IEnumerable<dynamic> records = sqlCon.Query("exec MySqlProcedure"); 

using (var writer = new StreamWriter("file.csv"))
using (var csv = new CsvWriter(writer, CultureInfo.InvariantCulture))
{
   var map = new CsvHelper.Configuration.DefaultClassMap();
   //ignore columns...
   csv.Context.RegisterClassMap(map);
   csv.WriteRecords(records);
}

[JoshClose]

You don't know the columns, but you do know ones you don't want?

void Main()
{
    var records = new List<dynamic>();
    
    {
        dynamic record = new ExpandoObject();
        record.Id = 1;
        record.Name = "one";
        record.DateTime = DateTimeOffset.Now;
        records.Add(record);
    }
        
    var config = new CsvConfiguration(CultureInfo.InvariantCulture)
    {
    };
    using (var writer = new StringWriter())
    using (var csv = new CsvWriter(writer, config))
    {
        for (var i = 0; i < records.Count; i++)
        {
            dynamic record = RemoveColumns(records[i]);
            
            if (i == 0)
            {
                csv.WriteDynamicHeader(records[0]);
                csv.NextRecord();
            }
            
            csv.WriteRecord(record);
            csv.NextRecord();
        }
        
        csv.Flush();
        writer.Dump();
    }
}

public dynamic RemoveColumns(dynamic record)
{
    var dictionary = (IDictionary<string, object>)record;
    // Remove any columns here you don't want.
    dictionary.Remove("DateTime");
    
    return dictionary;
}

[klaudyuxxx]

Thank you!
I was trying to go with class mapping thinking there's a simple way to create classes programmatically in runtime and add properties to it.
I finally went with a more straight forward approach and used SqlDataReader and CsvWriter.WriteField.
I'm curious if CsvHelper works better if it's used with class mapping versus how I did it below.
Does it cache internally the column types?

 Dictionary<string, string> columnsToExport = new Dictionary<string, string>()
            {
                {"id", "Code"},
                {"fooBar", "Foo Bar"}
            };

SqlCommand sqlCmd = new SqlCommand(sqlQuery, sqlCon);
SqlDataReader reader = sqlCmd.ExecuteReader();
using (var writer = new StreamWriter("sql.csv"))
using (var csv = new CsvWriter(writer, CultureInfo.InvariantCulture))
{
      var columnOrdinals = new List<int>();
      foreach (var coloanaDeExportat in columnsToExport)
      {
          try
          {
              var ordinal = reader.GetOrdinal(coloanaDeExportat.Key);
              //write the header columns
              csv.WriteField(coloanaDeExportat.Value);
              columnOrdinals.Add(ordinal);
          }
          catch (Exception e)
          {
              Console.WriteLine(e);
          }
      }

      if (columnOrdinals.Count == 0)
      {
          throw new Exception("No matching columns have been found withing SQL query: " + sqlQuery);
      }

      csv.NextRecord();
      while (reader.Read())
      {
          foreach (var columnOrdinal in columnOrdinals)
          {
              csv.WriteField(reader.GetValue(columnOrdinal));
          }
          csv.NextRecord();
      }
      csv.Flush();
      writer.Dump();
  }
}

[JoshClose]

Nope. All mappings do with WriteRecords is loop and call the WriteField methods for each property that's mapped. Unless you have a lot of properties, writing by hand using WriteField usually isn't much more code. WriteField will still do type conversions for you and you can pass your own ITypeConverter in too if you want.