关于糖尿病反应。这对我来说是可行的，除了我必须在乘法之前使用Floor来去除除法的小数部分。所以,

d = new DateTime((d.Ticks / TimeSpan.TicksPerSecond) * TimeSpan.TicksPerSecond);

就变成了

d = new DateTime(Math.Floor(d.Ticks / TimeSpan.TicksPerSecond) * TimeSpan.TicksPerSecond);

我本希望两个Long值的除法结果是Long，从而去掉小数部分，但它将其解析为Double，在乘法之后保留完全相同的值。

埃普西

一种简单的阅读方法是……

//Remove milliseconds
DateTime date = DateTime.Now;
date = DateTime.ParseExact(date.ToString("yyyy-MM-dd HH:mm:ss"), "yyyy-MM-dd HH:mm:ss", null);

和更多的……

//Remove seconds
DateTime date = DateTime.Now;
date = DateTime.ParseExact(date.ToString("yyyy-MM-dd HH:mm"), "yyyy-MM-dd HH:mm", null);

//Remove minutes
DateTime date = DateTime.Now;
date = DateTime.ParseExact(date.ToString("yyyy-MM-dd HH"), "yyyy-MM-dd HH", null);

//and go on...

我知道这很容易理解，但是它缺乏性能。

以下将适用于具有分数毫秒的DateTime，并保留Kind属性(Local, Utc或Undefined)。

DateTime dateTime = ... anything ...
dateTime = new DateTime(
    dateTime.Ticks - (dateTime.Ticks % TimeSpan.TicksPerSecond), 
    dateTime.Kind
    );

或等价的或更短的:

dateTime = dateTime.AddTicks( - (dateTime.Ticks % TimeSpan.TicksPerSecond));

这可以推广为一个扩展方法:

public static DateTime Truncate(this DateTime dateTime, TimeSpan timeSpan)
{
    if (timeSpan == TimeSpan.Zero) return dateTime; // Or could throw an ArgumentException
    if (dateTime == DateTime.MinValue || dateTime == DateTime.MaxValue) return dateTime; // do not modify "guard" values
    return dateTime.AddTicks(-(dateTime.Ticks % timeSpan.Ticks));
}

使用方法如下:

dateTime = dateTime.Truncate(TimeSpan.FromMilliseconds(1)); // Truncate to whole ms
dateTime = dateTime.Truncate(TimeSpan.FromSeconds(1)); // Truncate to whole second
dateTime = dateTime.Truncate(TimeSpan.FromMinutes(1)); // Truncate to whole minute
...

不那么明显，但快了2倍多:

// 10000000 runs

DateTime d = DateTime.Now;

// 484,375ms
d = new DateTime((d.Ticks / TimeSpan.TicksPerSecond) * TimeSpan.TicksPerSecond);

// 1296,875ms
d = d.AddMilliseconds(-d.Millisecond);

2上述解决方案的扩展方法

    public static bool LiesAfterIgnoringMilliseconds(this DateTime theDate, DateTime compareDate, DateTimeKind kind)
    {
        DateTime thisDate = new DateTime(theDate.Year, theDate.Month, theDate.Day, theDate.Hour, theDate.Minute, theDate.Second, kind);
        compareDate = new DateTime(compareDate.Year, compareDate.Month, compareDate.Day, compareDate.Hour, compareDate.Minute, compareDate.Second, kind);

        return thisDate > compareDate;
    }


    public static bool LiesAfterOrEqualsIgnoringMilliseconds(this DateTime theDate, DateTime compareDate, DateTimeKind kind)
    {
        DateTime thisDate = new DateTime(theDate.Year, theDate.Month, theDate.Day, theDate.Hour, theDate.Minute, theDate.Second, kind);
        compareDate = new DateTime(compareDate.Year, compareDate.Month, compareDate.Day, compareDate.Hour, compareDate.Minute, compareDate.Second, kind);

        return thisDate >= compareDate;
    }

用法:

bool liesAfter = myObject.DateProperty.LiesAfterOrEqualsIgnoringMilliseconds(startDateTime, DateTimeKind.Utc);

与其放弃毫秒然后比较，为什么不比较差异呢?

DateTime x; DateTime y;
bool areEqual = (x-y).TotalSeconds == 0;

or

TimeSpan precision = TimeSpan.FromSeconds(1);
bool areEqual = (x-y).Duration() < precision;