OceanBase STDDEV_SAMP

?STDDEV_SAMP? 函數(shù)計(jì)算樣本標(biāo)準(zhǔn)差。?STDDEV_SAMP? 函數(shù)將數(shù)值型數(shù)據(jù)作為參數(shù)，返回數(shù)值型數(shù)據(jù)。它與函數(shù) STDDEV 的不同之處在于，?STDDEV? 只有一行輸入數(shù)據(jù)時返回 0，而 ?STDDEV_SAMP? 返回 NULL。

說明 
樣本標(biāo)準(zhǔn)差是樣本方差的算術(shù)平方根。

語法

STDDEV_SAMP([ALL] expr) [ OVER (analytic_clause) ]

作為分析函數(shù)使用時，您需要使用窗口函數(shù)的完整語法，它對一組行的集合進(jìn)行計(jì)算并返回多個值 。作為聚合函數(shù)使用時，該函數(shù)對一組行的集合進(jìn)行聚合計(jì)算，結(jié)果只能返回一個值，此時不需要加 OVER 關(guān)鍵字。

參數(shù)

返回類型

返回類型與參數(shù) ?expr? 的數(shù)據(jù)類型相同。

示例

分析函數(shù)示例

以下語句創(chuàng)建了表 employees，并向里面插入數(shù)據(jù)：

CREATE TABLE employees (manager_id INT,last_name varchar(50),hiredate varchar(50),SALARY INT);
INSERT INTO employees VALUES(100, 'Raphaely', '2017-07-01', 1700);
INSERT INTO employees VALUES(100, 'De Haan', '2018-05-01',11000);      
INSERT INTO employees VALUES(100, 'Errazuriz', '2017-07-21', 1400);
INSERT INTO employees VALUES(100, 'Hartstein', '2019-05-01',14000);     
INSERT INTO employees VALUES(100, 'Raphaely', '2017-07-22', 1700);
INSERT INTO employees VALUES(100, 'Weiss',  '2019-07-11',13500);     
INSERT INTO employees VALUES(100, 'Russell', '2019-10-05', 13000);
INSERT INTO employees VALUES(100, 'Partners',  '2018-12-01',14000);     
INSERT INTO employees VALUES(200, 'Ross',  '2019-06-11',13500);     
INSERT INTO employees VALUES(200, 'Bell', '2019-05-25', 13000);
INSERT INTO employees VALUES(200, 'Part',  '2018-08-11',14000);
COMMIT;

調(diào)用函數(shù)并執(zhí)行以下語句：

SELECT manager_id, last_name, hiredate, salary,STDDEV_SAMP(salary) OVER (PARTITION BY manager_id
ORDER BY hiredate ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) AS cum_sdev 
FROM employees ORDER BY manager_id, last_name, hiredate, salary, cum_sdev;

返回結(jié)果如下：

+------------+-----------+------------+--------+-------------------------------------------+
| MANAGER_ID | LAST_NAME | HIREDATE   | SALARY | CUM_SDEV                                  |
+------------+-----------+------------+--------+-------------------------------------------+
|        100 | De Haan   | 2018-05-01 |  11000 | 4702.127178203498995615489088200868644482 |
|        100 | Errazuriz | 2017-07-21 |   1400 |  212.132034355964257320253308631454711785 |
|        100 | Hartstein | 2019-05-01 |  14000 | 6340.346993658943269176828928801701088079 |
|        100 | Partners  | 2018-12-01 |  14000 | 6064.899009876421676804205219406952308814 |
|        100 | Raphaely  | 2017-07-01 |   1700 |                                      NULL |
|        100 | Raphaely  | 2017-07-22 |   1700 |  173.205080756887729352744634150587236694 |
|        100 | Russell   | 2019-10-05 |  13000 | 6026.474330580265330900400184969999384459 |
|        100 | Weiss     | 2019-07-11 |  13500 | 6244.311697171159907069428668980211861012 |
|        200 | Bell      | 2019-05-25 |  13000 |  707.106781186547524400844362104849039285 |
|        200 | Part      | 2018-08-11 |  14000 |                                      NULL |
|        200 | Ross      | 2019-06-11 |  13500 |                                       500 |
+------------+-----------+------------+--------+-------------------------------------------+

聚合函數(shù)示例

調(diào)用函數(shù)并執(zhí)行以下語句：

SELECT STDDEV_SAMP(salary) FROM employees ;

查詢結(jié)果如下：

+-------------------------------------------+
| STDDEV_SAMP(SALARY)                       |
+-------------------------------------------+
| 5506.194858355615640082358245403620332764 |
+-------------------------------------------+