This CSV parser is a from-the-ground-up rewrite of the LumenWorks.Framework.IO parser. There were a few itches:

• First of all, I had a buggy CSV with double quotes inside double quoted fields. As I had no control over the input and it was virtually impossible to sanitize, I tried to modify the LumenWorks source code to fix this. However, I ended up rewriting the whole thing. The unit tests are actually the only code that was (almost) not touched :-)
• I wanted to add a more flexible way to convert string values to actual types.

There are a few static methods available in the ReadCsv class to easily read a CSV file using the DataReader style:

You need to add this using statement:

using Net.Code.Csv;

Now you can use code like this:

var myFileName = "path/to/csvfile";

using (var reader = ReadCsv.FromFile(myFileName)) {
    while (reader.Read()) {
        var record = new { 
            Name = reader["Name"];
            BirthDate = DateTime.Parse(reader["BirthDate"]);
        }
    }
}

You can also read from a stream or a string.

The example above assume some common defaults about the actual CSV layout and behaviour. You can of course change those through parameters.

Parameters that specify the CSV format include the encoding, the quote character, the field delimiter, a.o. Other parameters specify the 'behaviour' of the CSV reader: what to do with empty lines or missing fields, whether fields should be trimmed or not, etc.

Form more documentation, check the wiki