Sunday, July 13, 2008

LINQ to Text file through StreamReader Extension method


Text file data feeds come in different format but with common theme: parse a line into an object representation. Here is one example of the object:

public class BrokerDataItem : ILineParser<BrokerDataItem>
{
public string Ticker { get; set;}
public double Price { get; set;}
public long QTY { get; set;}
public string Trader { get; set;}
public DateTime TradeDate { get; set;}
public string TradeType { get; set;}

public BrokerDataItem() { }
public BrokerDataItem Parse(string[] p)
{
return null;
}
public BrokerDataItem Parse(string entireLine) {
string[] p = entireLine.Split(',');
Ticker = p[0].ToString();
Price = Convert.ToDouble(p[1]);
QTY = Convert.ToInt64(p[2]);
Trader = p[3].ToString();
TradeDate = Convert.ToDateTime(p[4]);
TradeType = p[5].ToString();
return this;
}

Note I abstracted parse operation into interface

public interface ILineParser<T>
{
T Parse(string[] p);
T Parse(string entireLine);
}

So that I can create an Extension Method for StreamReader that return the Data Object to LINQ. Consequently, this allows LINQ runs against data object of any source to get the required data elements:

public static class AnyThing
{
public static IEnumerable<T> GetData<T>(this StreamReader sr) where T : ILineParser<T>, new()
{
string L;
while ((L = sr.ReadLine()) != null)
{
T t = new T();
yield return t.Parse(L);
}
}
}


class Program
{
static void Main(string[] args)
{
StreamReader sr = new StreamReader(@"..\..\Data.txt");

var t =
from data in sr.GetData<BrokerDataItem>()
select data;
foreach (BrokerDataItem i in t)
Console.WriteLine(i.Ticker+" " + i.Price.ToString());
Console.ReadLine();
}
}

Here are the data:
FBN U8,134.79, 1000,JIMMYD,7/14/2008,BUYL
FBN U8,444.90, 7832,JIMMYD,7/14/2008,BUYL
FBN U8,4567.9, 1000,JIMMYD,7/14/2008,SELL
FBN U8,581.79, 5634,JIMMYD,7/14/2008,BUYL
FBN U8,724.78, 1212,JIMMYD,7/14/2008,SELL
FBN U8,932.01, 8892,JIMMYD,7/14/2008,SELL
FBN U8,678.88, 3,JIMMYD,7/14/2008,BUYL

Finally, extension method is not required since you can just create an IEnumerable and use it in LINQ.

No comments: