现有表结构：

CREATE TABLE `example_dataset` (

`id` int(11) unsigned NOT NULL AUTO_INCREMENT,

`tel` bigint(11) DEFAULT NULL,

`gender` varchar(11) DEFAULT NULL,

PRIMARY KEY (`id`)

) ENGINE=InnoDB AUTO_INCREMENT=1 DEFAULT CHARSET=utf8mb4;

插入数据

INSERT INTO `example_dataset` (`id`, `tel`, `gender`)

VALUES

(2, 7136609221, 'male'), --'male'是男，'female'是女

(3, 7136609222, 'male'), (6, 7136609222, 'male'), (4, 7136609223, 'female'), (7, 7136609223, 'female'), (5, 7136609228, 'male');

现在sql记录里面有重复的。我想统计表中有多少人，多少男的，多少女的。

SELECT COUNT(DISTINCT tel, gender) as gender_count,

COUNT(DISTINCT tel, gender = 'male') as man_count,

SUM(if(gender = 'female', 1, 0)) as woman_count

FROM example_dataset;

显然结果是不对的。

SELECT COUNT(*) gender_count,

SUM(IF(gender='male',1,0)) male_count,

SUM(IF(gender='female',1,0)) female_count

FROM (

SELECT DISTINCT tel, gender

FROM example_dataset

) t

这样结果是对的。但是我想如果example_dataset表的数据量比较大，子查询会影响效率。

最终答案：

SELECT COUNT(DISTINCT tel) gender_count,

COUNT(DISTINCT CASE WHEN gender = 'male' THEN tel END) male_count,

COUNT(DISTINCT CASE WHEN gender = 'female' THEN tel END) female_count

FROM example_dataset