全部产品
Search
文档中心

云原生大数据计算服务 MaxCompute:DATEDIFF

更新时间:Jun 15, 2023

计算两个时间date1date2的差值,将差值以指定的时间单位datepart表示。

命令格式

bigint datediff(date|datetime|timestamp <date1>, date|datetime|timestamp <date2>, string <datepart>)

参数说明

  • date1date2:必填。DATE、DATETIME或TIMESTAMP类型。被减数和减数。如果输入为STRING类型,且MaxCompute项目的数据类型版本是1.0,则会隐式转换为DATETIME类型后参与运算。

  • datepart:可选。时间单位,STRING类型常量。

    如果您打开了MaxCompute新数据类型开关,可以不填写datepart参数,默认日期格式为天。该字段也支持扩展的日期格式:年-year、月-month-mon、日-day和小时-hour

    说明

    系统计算时会按照datepart切掉低单位部分,然后再计算结果。

返回值说明

返回BIGINT类型。返回规则如下:

  • date1date2非DATE、DATETIME或TIMESTAMP类型时,返回报错。

  • 如果date1小于date2,返回值为负数。

  • date1date2值为NULL时,返回NULL。

  • datepart值为NULL时,返回NULL。

示例数据

为便于理解各函数的使用方法,本文为您提供源数据,基于源数据提供函数相关示例。创建表mf_date_fun_t,并添加数据,命令示例如下。

create table if not exists mf_date_fun_t(
    id      int,
    date1   date,
    datetime1   datetime,
    timestamp1 timestamp,
    date2   date,
    datetime2   datetime,
    timestamp2 timestamp,
    date3 string,
    date4 bigint);
insert into mf_date_fun_t values
(1,DATE'2021-11-29',DATETIME'2021-11-29 00:01:00',TIMESTAMP'2021-01-11 00:00:00.123456789',DATE'2021-10-29',DATETIME'2021-10-29 00:00:00',TIMESTAMP'2021-10-11 00:00:00.123456789','2021-11-20',123456780),
(2,DATE'2021-11-28',DATETIME'2021-11-28 00:02:00',TIMESTAMP'2021-02-11 00:00:00.123456789',DATE'2021-10-29',DATETIME'2021-10-29 00:00:00',TIMESTAMP'2021-10-11 00:00:00.123456789','2021-11-21',123456781),
(3,DATE'2021-11-27',DATETIME'2021-11-27 00:03:00',TIMESTAMP'2021-03-11 00:00:00.123456789',DATE'2021-10-29',DATETIME'2021-10-29 00:00:00',TIMESTAMP'2021-10-11 00:00:00.123456789','2021-11-22',123456782),
(4,DATE'2021-11-26',DATETIME'2021-11-26 00:04:00',TIMESTAMP'2021-04-11 00:00:00.123456789',DATE'2021-10-29',DATETIME'2021-10-29 00:00:00',TIMESTAMP'2021-10-11 00:00:00.123456789','2021-11-23',123456783),
(5,DATE'2021-11-25',DATETIME'2021-11-25 00:05:00',TIMESTAMP'2021-05-11 00:00:00.123456789',DATE'2021-10-29',DATETIME'2021-10-29 00:00:00',TIMESTAMP'2021-10-11 00:00:00.123456789','2021-11-24',123456784),
(6,DATE'2021-11-24',DATETIME'2021-11-24 00:06:00',TIMESTAMP'2021-06-11 00:00:00.123456789',DATE'2021-10-29',DATETIME'2021-10-29 00:00:00',TIMESTAMP'2021-10-11 00:00:00.123456789','2021-11-25',123456785),
(7,DATE'2021-11-23',DATETIME'2021-11-23 00:07:00',TIMESTAMP'2021-07-11 00:00:00.123456789',DATE'2021-10-29',DATETIME'2021-10-29 00:00:00',TIMESTAMP'2021-10-11 00:00:00.123456789','2021-11-26',123456786),
(8,DATE'2021-11-22',DATETIME'2021-11-22 00:08:00',TIMESTAMP'2021-08-11 00:00:00.123456789',DATE'2021-10-29',DATETIME'2021-10-29 00:00:00',TIMESTAMP'2021-10-11 00:00:00.123456789','2021-11-27',123456787),
(9,DATE'2021-11-21',DATETIME'2021-11-21 00:09:00',TIMESTAMP'2021-09-11 00:00:00.123456789',DATE'2021-10-29',DATETIME'2021-10-29 00:00:00',TIMESTAMP'2021-10-11 00:00:00.123456789','2021-11-28',123456788),
(10,DATE'2021-11-20',DATETIME'2021-11-20 00:10:00',TIMESTAMP'2021-10-11 00:00:00.123456789',DATE'2021-10-29',DATETIME'2021-10-29 00:00:00',TIMESTAMP'2021-10-11 00:00:00.123456789','2021-11-29',123456789);

查询表mf_date_fun_t中的数据,命令示例如下:

select * from mf_date_fun_t;
--返回结果。
+------------+------------+---------------------+-------------------------+------------+---------------------+-------------------------+------------+------------+
| id         | date1      | datetime1           | timestamp1              | date2      | datetime2           | timestamp2              | date3      | date4      |
+------------+------------+---------------------+-------------------------+------------+---------------------+-------------------------+------------+------------+
| 1          | 2021-11-29 | 2021-11-29 00:01:00 | 2021-01-11 00:00:00.123 | 2021-10-29 | 2021-10-29 00:00:00 | 2021-10-11 00:00:00.123 | 2021-11-20 | 123456780  |
| 2          | 2021-11-28 | 2021-11-28 00:02:00 | 2021-02-11 00:00:00.123 | 2021-10-29 | 2021-10-29 00:00:00 | 2021-10-11 00:00:00.123 | 2021-11-21 | 123456781  |
| 3          | 2021-11-27 | 2021-11-27 00:03:00 | 2021-03-11 00:00:00.123 | 2021-10-29 | 2021-10-29 00:00:00 | 2021-10-11 00:00:00.123 | 2021-11-22 | 123456782  |
| 4          | 2021-11-26 | 2021-11-26 00:04:00 | 2021-04-11 00:00:00.123 | 2021-10-29 | 2021-10-29 00:00:00 | 2021-10-11 00:00:00.123 | 2021-11-23 | 123456783  |
| 5          | 2021-11-25 | 2021-11-25 00:05:00 | 2021-05-11 00:00:00.123 | 2021-10-29 | 2021-10-29 00:00:00 | 2021-10-11 00:00:00.123 | 2021-11-24 | 123456784  |
| 6          | 2021-11-24 | 2021-11-24 00:06:00 | 2021-06-11 00:00:00.123 | 2021-10-29 | 2021-10-29 00:00:00 | 2021-10-11 00:00:00.123 | 2021-11-25 | 123456785  |
| 7          | 2021-11-23 | 2021-11-23 00:07:00 | 2021-07-11 00:00:00.123 | 2021-10-29 | 2021-10-29 00:00:00 | 2021-10-11 00:00:00.123 | 2021-11-26 | 123456786  |
| 8          | 2021-11-22 | 2021-11-22 00:08:00 | 2021-08-11 00:00:00.123 | 2021-10-29 | 2021-10-29 00:00:00 | 2021-10-11 00:00:00.123 | 2021-11-27 | 123456787  |
| 9          | 2021-11-21 | 2021-11-21 00:09:00 | 2021-09-11 00:00:00.123 | 2021-10-29 | 2021-10-29 00:00:00 | 2021-10-11 00:00:00.123 | 2021-11-28 | 123456788  |
| 10         | 2021-11-20 | 2021-11-20 00:10:00 | 2021-10-11 00:00:00.123 | 2021-10-29 | 2021-10-29 00:00:00 | 2021-10-11 00:00:00.123 | 2021-11-29 | 123456789  |
+------------+------------+---------------------+-------------------------+------------+---------------------+-------------------------+------------+------------+

使用示例:静态数据示例

--假设start=2005-12-31 23:59:59,end=2006-01-01 00:00:00,则:
    --返回1。
    select datediff(end, start, 'dd'); 
    --返回1。
    select datediff(end, start, 'mm');
    --返回1。
    select datediff(end, start, 'yyyy');
    --返回1。
    select datediff(end, start, 'hh');
    --返回1。
    select datediff(end, start, 'mi');
    --返回1。
    select datediff(end, start, 'ss');
--返回1800。
select datediff(datetime'2013-05-31 13:00:00', datetime'2013-05-31 12:30:00', 'ss');
--返回30。
set odps.sql.type.system.odps2=false; 
select datediff('2013-05-31 13:00:00', '2013-05-31 12:30:00', 'mi');
--返回11。
select datediff(date '2013-05-21', date '2013-05-10', 'dd');
--开启2.0新类型。此命令需要与SQL语句一起提交。返回1800。
set odps.sql.type.system.odps2=true;
select datediff(timestamp '2013-05-31 13:00:00', timestamp '2013-05-31 12:30:00', 'ss');
--假设start = 2018-06-04 19:33:23.234,end = 2018-06-04 19:33:23.250,含毫秒的日期不属于标准DATETIME式样,不能直接隐式转换,此处需进行显示转换。返回16。
select datediff(to_date('2018-06-04 19:33:23.250', 'yyyy-mm-dd hh:mi:ss.ff3'), to_date('2018-06-04 19:33:23.234', 'yyyy-mm-dd hh:mi:ss.ff3') , 'ff3');
--返回NULL。
select datediff(date '2013-05-21', date '2013-05-10', null);
--返回NULL。
select datediff(date '2013-05-21', null, 'dd');

使用示例:表数据示例

基于示例数据,计算date1和date2、datetime1和datetime2、timestamp1和timestamp2列的差值,并以指定单位表示,命令示例如下。

--开启2.0新类型。此命令需要与SQL语句一起提交。
set odps.sql.type.system.odps2=true;
select date1, date2, datediff(date1,date2,'dd') as date1_date2_datediff, datetime1, datetime2, datediff(datetime1, datetime2,'dd') as datetime1_datetime2_datediff, timestamp1, timestamp2, datediff(timestamp1, timestamp2,'mm') as timestamp1_timestamp2_datediff from mf_date_fun_t;

返回结果如下。

+------------+------------+----------------------+---------------------+---------------------+------------------------------+-------------------------------+-------------------------------+--------------------------------+
| date1      | date2      | date1_date2_datediff | datetime1           | datetime2           | datetime1_datetime2_datediff | timestamp1                    | timestamp2                    | timestamp1_timestamp2_datediff |
+------------+------------+----------------------+---------------------+---------------------+------------------------------+-------------------------------+-------------------------------+--------------------------------+
| 2021-11-29 | 2021-10-29 | 31                   | 2021-11-29 00:01:00 | 2021-10-29 00:00:00 | 31                           | 2021-01-11 00:00:00.123456789 | 2021-10-11 00:00:00.123456789 | -9                             |
| 2021-11-28 | 2021-10-29 | 30                   | 2021-11-28 00:02:00 | 2021-10-29 00:00:00 | 30                           | 2021-02-11 00:00:00.123456789 | 2021-10-11 00:00:00.123456789 | -8                             |
| 2021-11-27 | 2021-10-29 | 29                   | 2021-11-27 00:03:00 | 2021-10-29 00:00:00 | 29                           | 2021-03-11 00:00:00.123456789 | 2021-10-11 00:00:00.123456789 | -7                             |
| 2021-11-26 | 2021-10-29 | 28                   | 2021-11-26 00:04:00 | 2021-10-29 00:00:00 | 28                           | 2021-04-11 00:00:00.123456789 | 2021-10-11 00:00:00.123456789 | -6                             |
| 2021-11-25 | 2021-10-29 | 27                   | 2021-11-25 00:05:00 | 2021-10-29 00:00:00 | 27                           | 2021-05-11 00:00:00.123456789 | 2021-10-11 00:00:00.123456789 | -5                             |
| 2021-11-24 | 2021-10-29 | 26                   | 2021-11-24 00:06:00 | 2021-10-29 00:00:00 | 26                           | 2021-06-11 00:00:00.123456789 | 2021-10-11 00:00:00.123456789 | -4                             |
| 2021-11-23 | 2021-10-29 | 25                   | 2021-11-23 00:07:00 | 2021-10-29 00:00:00 | 25                           | 2021-07-11 00:00:00.123456789 | 2021-10-11 00:00:00.123456789 | -3                             |
| 2021-11-22 | 2021-10-29 | 24                   | 2021-11-22 00:08:00 | 2021-10-29 00:00:00 | 24                           | 2021-08-11 00:00:00.123456789 | 2021-10-11 00:00:00.123456789 | -2                             |
| 2021-11-21 | 2021-10-29 | 23                   | 2021-11-21 00:09:00 | 2021-10-29 00:00:00 | 23                           | 2021-09-11 00:00:00.123456789 | 2021-10-11 00:00:00.123456789 | -1                             |
| 2021-11-20 | 2021-10-29 | 22                   | 2021-11-20 00:10:00 | 2021-10-29 00:00:00 | 22                           | 2021-10-11 00:00:00.123456789 | 2021-10-11 00:00:00.123456789 | 0                              |
+------------+------------+----------------------+---------------------+---------------------+------------------------------+-------------------------------+-------------------------------+--------------------------------+

相关函数

DATEDIFF函数属于日期函数,更多日期计算、日期转换的相关函数请参见日期函数