This is really a regex question.

I am wonding if anyone knows a good Regex expression that would pull a valid
date from a string.

I have used:

            strValue = Regex.Replace(valueIn, @"[^\d/]", "");

which works most of the time.

But I have some cases where I have strings like:

05/07/08(-4%)
09/19/08 DOM 55
09/19/2008 DOM 53
FOR 09/15/08 -23

Thanks,

Tom

What is really strange is that 05/07/08(-4%) is converting to:

05/07/084

Convert.ToDateTime is change this to: 05/07/0084 12:00:00 AM????

But Sql says it is invalid and gives me an error:

   Message: SqlDateTime overflow. Must be between 1/1/1753 12:00:00 AM and
12/31/9999 11:59:59 PM.

It turns out all of these examples die in SQL but pass in VS.

The code I use is:

        public static DateTime? CleanDates(string valueIn)
        {
            DateTime? datValue = null;
            string strValue;

            strValue = Regex.Replace(valueIn, @"[^\d/]", "");
            try
            {
                datValue = Convert.ToDateTime(strValue);
            }
            catch
            {
                datValue = null;
            }
            return datValue;
        }

I would have assumed it would take the Catch, but doesn't.

Why is this?

Thanks,

Tom

Show quoteHide quote

tshad wrote:
SQLServer DATETIME apparently has a different range than
..NET System.DateTime - that is not particular surprising.

Arne

Show quote Hide quote I agree.  I don't know what the range is, but that wouldn't help anyway.  It
can be a valid date and not be THE valid date - which this showed.  I wasn't
expecting this - but should have.

Thanks,

Tom
Show quoteHide quote

tshad wrote:
Show quoteHide quotehttp://msdn.microsoft.com/en-us/library/system.datetime.aspx

The DateTime value type represents dates and times with values ranging
from 12:00:00 midnight, January 1, 0001 Anno Domini (Common Era) through
11:59:59 P.M., December 31, 9999 A.D. (C.E.)

Time values are measured in 100-nanosecond units called ticks, and a
particular date is the number of ticks since 12:00 midnight, January 1,
0001 A.D. (C.E.) in the GregorianCalendar calendar.

http://msdn.microsoft.com/en-us/library/ms187819.aspx

Date range         January 1, 1753, through December 31, 9999

Time range         00:00:00 through 23:59:59.997

Accuracy           Rounded to increments of .000, .003, or .007 seconds

Arne

On Fri, 19 Dec 2008 17:26:41 -0800, tshad <t**@dslextreme.com> wrote:

Why is that strange?  Your regex basically removes anything that's not a 
digit or a '/' character.

Well, what did you expect it to do with the year "084"?

I don't know much about SQL, but it looks like the range of valid dates in 
SQL does not include the year 84.  Not that I see any reason you'd _want_ 
it to accept the year 84, given the original string you started with.

What does that mean?  It's clear why the result of your conversion isn't 
accepted by SQL.  But what does it mean for something to "pass in VS"?

Show quoteHide quoteWhy would you have assumed that the execution would wind up in the "catch" 
clause?  You had a successful conversion from the string to a DateTime 
value.

As far as the original question goes, you're not being specific enough 
about what it is you really want to do.  It seems to me that if you want 
to actually extract dates of the form "mm/dd/yy" or "mm/dd/yyyy", you 
should write a regex pattern that matches _that_, rather than trying to 
remove all the characters from the string that aren't digits or '/' 
characters.

From your examples, it appears that you always have two-digit months and 
dates, but may have two- or four-digit years.  So, perhaps rather than 
using the Replace() method, you could use the Match() method, with a 
string like this: @"\d\d/\d\d/\d\d(\d\d)?"  Keeping in mind that I'm no 
regex expert, and might have something not quite right in that one.  But 
hopefully you get the idea; that is, that there is in fact a specific 
pattern that you're looking for, so matching against that pattern is 
likely to be more fruitful than trying to exclude specific characters.

Pete

Show quote Hide quote I know that.  I didn't know the range would be that much different from SQL
but it obviously is.

I don't want it to accept this date.  The problem is that my regex only gets
rid of everything that is not a number and a "/".  This works for about 99%
of my dates (out of about 200,000 at the moment).

But I would like to figure out a better regex option.

I mean that it considers it a valid date - which it is but not really THE
valid date.

Show quoteHide quoteI agree and that is my problem.  It passes (valid) in VS and then dies in
SQL.  But SQL is only dropping it because it is outside the date range.
That doesn't necessarily making it correct.  Being a valid date doesn't make
it the correct date.  Which is my problem.
I agree.  Which was what I was asking.  I just wanted to see if anyone knew
of one off hand.
I agree.  I am actually looking for 1 or 2 characters in the days and months
and 2 or 4 in the last.  Exactly that pattern.  Anything else I want to get
rid of.

Thanks,

Tom
Show quoteHide quote

tshad wrote:
Show quoteHide quoteFor inspiration see below.

Arne

=======================

using System;
using System.Globalization;
using System.Text.RegularExpressions;

namespace E
{
     public class Program
     {
         private static readonly Regex re = new
Regex(@"\d{2}/\d{2}/\d{2}", RegexOptions.Compiled);
         public static DateTime Parse(string s)
         {
             return DateTime.Parse(re.Match(s).Value, new
CultureInfo("en-US"));
         }
         public static void Main(string[] args)
         {
             Console.WriteLine(Parse("05/07/08(-4%)"));
             Console.WriteLine(Parse("09/19/08 DOM 55"));
             Console.WriteLine(Parse("09/19/2008 DOM 53"));
             Console.WriteLine(Parse("FOR 09/15/08 -23"));
             Console.ReadKey();
         }
     }
}

Show quote Hide quote That looks good.  Would that handle 1 and 2 digits for days and months and 2
and 4 for years?

Thanks,

Tom
Show quoteHide quote

tshad wrote:
Show quoteHide quote@"\d{1,2}/\d{1,2}/\d{2,4}"

is more flexible.

Arne

Inheritance problem
Multi threaded design patterns
Mapping
Combobox doesn't work if the form has TopMost = true
how to convert a byte array with string,int and double embeded?
Inherit Overrides
Linq.Table<variable> is it possible
Do SingleCall server activation objects retain state across calls?
Linq To Sql: Contains any.
Class and Methods access