Question

Anyone know how to (or if its possible) when using linq if one of the iterations fails to just ignore the fails and keep the successful. Example: when I’m importing a CSV:

string[] csvRows = File.ReadAllLines(fiList[fileCount].FullName);      

var Rows = from row in csvRows.Skip(1)
           let data = row.Split(',')
           select new
           {
               Reference = data[0],
               RequestDate = Convert.ToDateTime(data[1]),
           };

and the csv looks like:

Reference, RequestDate ref1,20/06/10
00:00:00 ref2,55/06/10 00:00:00

Obviously, it’ll fail on the second one as it can’t convert to a datetime (“String was not recognized as a valid DateTime.”). Is there any way to force .net to ignore this and just return the first row?

Cheers in advance

Stu

Answer 1

You could test and filter out the invalid rows:

var rows = csvRows.Skip(1).Select(r => 
    {
        var data = r.Split(',');
        DateTime result;

        if(DateTime.TryParse(data[1], out result))
            return new { Reference = data[0], RequestDate = result};
        else return null;
    }).Where( x=> x!=null);

Personally I’d go for a foreach loop in this case which would make the code more readable.

An alternative to support readability is to factor out the validity testing (just an example):

public static bool IsValidColumnSet(string[] data)
{
    DateTime result;

    if(DateTime.TryParse(data[1], out result))
        return true;
    else return false;
}

var Rows = from row in csvRows.Skip(1)
           let data = row.Split(',')
           where IsValidColumnSet(data)
           select new
           {
               Reference = data[0],
               RequestDate = Convert.ToDateTime(data[1]),
           };