Question

I was greeted with a nasty bug today. The task is pretty trivial, all I needed to do is to convert the DateTime object to string in “yyyymmdd” format. The “yyyymmdd” part was stated in the development doc from the external software vendor. So, I conveniently copied the string from their file and pasted to my code. So I got the next

public string GetDateString(DateTime dateTime)
{
    return dateTime.ToString("yyyymmdd");
}

Pretty simple. So simple that I didn’t feel like to unit test the method. 20 minutes later, when other parts of my component are done. I started the app to check if things went right. Almost immediately I notice some supposed-to-be date field in my web page is displaying 20091511! This can’t be right, there is no 15th month of a year. So, I rushed back to my code to check possible errors. It turns out the that I should have used “yyyyMMdd” instead of “yyyymmdd” when converting DateTime to string.

Admitted, this bug was due to my lack of attention to details. The difference between “mm” and “MM” are cleared stated in all C# references. I still would like to argue that it’s pretty easy to overlook the differences if one doesn’t work with this kind of tasks everyday.

My question is: Is there a clean(i.e. no magic string) way to do the coverings in one line of code? Thereturn dateTime.Year + "" + dateTime.Month + "" + dateTime.Day; code seems to be working but it’s too much like hacking.

Update: Looks like the string format way is the best C# can offer. Maybe I am being brain washed, but I still think this kind of programming style belongs to low-level languages such as c.

Answer 1

String.Format("{0:0000}{1:00}{2:00}", dateTime.Year, dateTime.Month, dateTime.Day);

You could use this instead, I prefer the terse format though. Instead of 00 you can also use MM for specific month formatting (like in DateTime.ToString()).