Trouble parsing dates

[Arokh]

Hi,

I need to parse a specific date format, as fast as possible,

since there are around a million of them I need to parse in a go.

 

The date format looks like this:

"7/30 17:15:39.140"

"7/30 10:07:03.156"

 

After searching for an example on the MSDN I found this:

http://msdn.microsoft.com/en-us/library/w2sa9yss.aspx

http://msdn.microsoft.com/en-us/library/8kb3ddd4.aspx

 

[CSharp]

string dateString, format;

DateTime result;

CultureInfo provider = CultureInfo.InvariantCulture;

 

// Parse date and time with custom specifier.

dateString = "Sun 15 Jun 2008 8:30 AM -06:00";

format = "ddd dd MMM yyyy h:mm tt zzz";

try {

result = DateTime.ParseExact(dateString, format, provider);

Console.WriteLine("{0} converts to {1}.", dateString, result.ToString());

} catch(FormatException) {

Console.WriteLine("{0} is not in the correct format.", dateString);

}

 

[/CSharp]

 

So I tried to adapt to my needs but don't seem to get it right.

I always get an 'System.FormatException'

 

Here is what I tried:

[CSharp]

string dateString, format;

DateTime result;

CultureInfo provider = CultureInfo.InvariantCulture;

 

// Parse date and time with custom specifier.

dateString = "2008/7/30 13:52:15.765";

format = "yyyy/MM/DD HH:mm:ss.FFF";

try {

result = DateTime.ParseExact(dateString, format, provider);

Console.WriteLine("{0} converts to {1}.", dateString, result.ToString());

} catch(FormatException) {

Console.WriteLine("{0} is not in the correct format.", dateString);

}

[/CSharp]

 

On another note.

Is this faster than:

[CSharp]

string[] TimeStampParts = logEntry.Split(new Char[] { '/', ' ', ':', '.' });

TimeStamp = new DateTime(

(long.Parse(TimeStampParts[5]) +

(long.Parse(TimeStampParts[4]) +

(long.Parse(TimeStampParts[3]) +

(long.Parse(TimeStampParts[2]) +

(long.Parse(TimeStampParts[1]) +

long.Parse(TimeStampParts[0]) * 30) * 24) * 60) * 60) * 1000) * 10000);[/CSharp]

?

 

If not, are there any methods/tricks I can use to make it faster?

[PlausiblyDamp]

A format of "yyyy/MM/dd HH:mm:ss.fff" seems to work fine though, not looked into the difference between fff and FFF though so you may want to see if this is suitable.