GROUP BY DATEADD(day, DATEDIFF(day, 0, MyDateTimeColumn), 0)

或者在SQL Server 2008以后，你可以像@Oded建议的那样简单地转换为Date:

GROUP BY CAST(orderDate AS DATE)

将值转换为组的日期类型。

GROUP BY CAST(myDateTime AS DATE)

在Sql 2008之前，通过去掉日期部分:

GROUP BY CONVERT(CHAR(8),DateTimeColumn,10)

下面是我使用的一个例子，当我需要计算特定日期的记录数量而不包括时间部分时:

select count(convert(CHAR(10), dtcreatedate, 103) ),convert(char(10), dtcreatedate, 103)
FROM dbo.tbltobecounted
GROUP BY CONVERT(CHAR(10),dtcreatedate,103)
ORDER BY CONVERT(CHAR(10),dtcreatedate,103)

下面的例子在oracle中工作得很好

select to_char(columnname, 'DD/MON/yyyy'), count(*) from table_name group by to_char(createddate, 'DD/MON/yyyy');

GROUP BY DATE(date_time_column)

我相信你需要在每年的那个月的那一天进行分组。 那么为什么不使用TRUNK_DATE函数呢? 它的工作原理如下:

Group By DATE_TRUNC('day' , 'occurred_at_time')

CAST datetime字段到日期

select  CAST(datetime_field as DATE), count(*) as count from table group by CAST(datetime_field as DATE);

好吧，对我来说，它几乎是直接的，我使用cast和groupby:

例子:

Select cast(created_at as date), count(1) from dbname.tablename GROUP BY cast(created_at as date)

注意:我在MSSQL 2016上使用这个。