MySQL必知必会07：条件语句where与having不同
阅读整理自《MySQL 必知必会》- 朱晓峰，详细内容请登录 极客时间 官网购买专栏。
文章目录
-
- 实际查询需求
- where
- having
-
- group by
- having
- 正确使用where和having
查询的时候，经常需要按条件对查询结果进行筛选，这就要用到条件语句 WHERE 和 HAVING 了。
-
WHERE 是直接对表中的字段进行限定，来筛选结果；
-
HAVING 则需要跟分组关键字 GROUP BY 一起使用，通过对分组字段或分组计算函数进行限定，来筛选结果。
虽然它们都是对查询进行限定，却有着各自的特点和适用场景。很多时候，会遇到 2 个都可以用的情况。一旦用错，就很容易出现执行效率低下、查询结果错误，甚至是查询无法运行的情况。
实际查询需求
超市的经营者提出，要查单笔销售金额超过 50 元的商品。我们来分析一下这个需求：需要查询出一个商品记录集，限定条件是单笔销售金额超过 50 元。
假设我们有一个这样的商品信息表（demo.goodsmaster），里面有 2 种商品：书和笔。
数据准备：
show databases;
use demo;
create table demo.goodsmaster
(
itemnumber int primary key auto_increment,
barcode text,
goodsname text,
specification text,
unit text,
saleprice decimal(5,2)
);
describe goodsmaster;
insert into demo.goodsmaster (itemnumber, barcode, goodsname, specification, unit, saleprice) values (1, '0001', '书', '', '本', 89);
insert into demo.goodsmaster (itemnumber, barcode, goodsname, specification, unit, saleprice) values (2, '0002', '笔', '', '支', 5);
select * from demo.goodsmaster;
create table demo.transactiondetails
(
transactionid int,
itemnumber int,
quantity decimal(10,3),
price decimal(5,2),
salesvalue decimal(5,2)
);
describe transactiondetails;
insert into demo.transactiondetails (transactionid, itemnumber, quantity, price, salesvalue) values (1, 1, 1, 89, 89);
insert into demo.transactiondetails (transactionid, itemnumber, quantity, price, salesvalue) values (1, 2, 2, 5, 10);
insert into demo.transactiondetails (transactionid, itemnumber, quantity, price, salesvalue) values (2, 1, 2, 89, 178);
insert into demo.transactiondetails (transactionid, itemnumber, quantity, price, salesvalue) values (3, 2, 10, 5, 50);
select * from demo.transactiondetails;
商品信息表：
mysql> select * from demo.goodsmaster;
+------------+---------+-----------+---------------+------+-----------+
| itemnumber | barcode | goodsname | specification | unit | saleprice |
+------------+---------+-----------+---------------+------+-----------+
| 1 | 0001 | 书 | | 本 | 89.00 |
| 2 | 0002 | 笔 | | 支 | 5.00 |
+------------+---------+-----------+---------------+------+-----------+
2 rows in set (0.01 sec)
商品销售明细表：
mysql> select * from demo.transactiondetails;
+---------------+------------+----------+-------+------------+
| transactionid | itemnumber | quantity | price | salesvalue |
+---------------+------------+----------+-------+------------+
| 1 | 1 | 1.000 | 89.00 | 89.00 |
| 1 | 2 | 2.000 | 5.00 | 10.00 |
| 2 | 1 | 2.000 | 89.00 | 178.00 |
| 3 | 2 | 10.000 | 5.00 | 50.00 |
+---------------+------------+----------+-------+------------+
4 rows in set (0.00 sec)
使用 where查询：
select distinct b.goodsname
from demo.transactiondetails as a
join demo.goodsmaster as b
on (a.itemnumber = b.itemnumber)
where a.salesvalue > 50;
使用 having 查询：
select b.goodsname
from demo.transactiondetails as a
join demo.goodsmaster as b
on (a.itemnumber = b.itemnumber)
group by b.goodsname
having max(a.salesvalue) > 50;
结果都是：
+-----------+
| goodsname |
+-----------+
| 书 |
+-----------+
1 row in set (0.02 sec)
where
分析一下使用 WHERE 条件的查询语句，是如何执行这个查询的
-
首先，MySQL 从数据表
demo.transactiondetails中抽取满足条件a.salesvalue>50的记录：mysql> select * from demo.transactiondetails as a where a.salesvalue > 50; +---------------+------------+----------+-------+------------+ | transactionid | itemnumber | quantity | price | salesvalue | +---------------+------------+----------+-------+------------+ | 1 | 1 | 1.000 | 89.00 | 89.00 | | 2 | 1 | 2.000 | 89.00 | 178.00 | +---------------+------------+----------+-------+------------+ 2 rows in set (0.00 sec) -
为了获取到销售信息所对应的商品名称，需要通过公共字段
itemnumbers与数据表demo.goodsmaster进行关联，从demo.goodsmaster中获取商品名称：mysql> select a.*, b.goodsname -> from demo.transactiondetails a -> join demo.goodsmaster b -> on (a.itemnumber = b.itemnumber) -> where a.salesvalue > 50; +---------------+------------+----------+-------+------------+-----------+ | transactionid | itemnumber | quantity | price | salesvalue | goodsname | +---------------+------------+----------+-------+------------+-----------+ | 1 | 1 | 1.000 | 89.00 | 89.00 | 书 | | 2 | 1 | 2.000 | 89.00 | 178.00 | 书 | +---------------+------------+----------+-------+------------+-----------+ 2 rows in set (0.00 sec) -
查询商品名称，就会出现两个重复的记录：
mysql> select b.goodsname from demo.transactiondetails a join demo.goodsmaster b on (a.itemnumber = b.itemnumber) where a.salesvalue > 50; +-----------+ | goodsname | +-----------+ | 书 | | 书 | +-----------+ 2 rows in set (0.00 sec) -
为了消除重复的语句，这里我们需要用到一个关键字：
DISTINCT，它的作用是返回唯一不同的值。比如，DISTINCT 字段 1，就表示返回所有字段 1 的不同的值mysql> select distinct( b.goodsname )from demo.transactiondetails a join demo.goodsmaster b on (a.itemnumber = b.itemnumber) where a.salesvalue > 50; +-----------+ | goodsname | +-----------+ | 书 | +-----------+ 1 row in set (0.00 sec)
WHERE 关键字的特点是，直接用表的字段对数据集进行筛选。
如果需要通过关联查询从其他的表获取需要的信息，那么执行的时候，也是先通过 WHERE 条件进行筛选，用筛选后的比较小的数据集进行连接。
这样一来，连接过程中占用的资源比较少，执行效率也比较高。
having
HAVING 不能单独使用，必须要跟 GROUP BY 一起使用。可以把 GROUP BY 理解成对数据进行分组，方便对组内的数据进行统计计算。
group by
先说明 GROUP BY 如何使用，以及如何在分组里面进行统计计算。
假设现在有一组销售数据，我们需要从里面查询每天、每个收银员的销售数量和销售金额。
数据准备：
create table demo.transactionhead
(
transactionid int primary key auto_increment,
transactionno text,
operatorid int,
transdate datetime
);
insert into demo.transactionhead (transactionid, transactionno, operatorid, transdate) values (1, '0120201201000001', 1, '2020-12-10 00:00:00');
insert into demo.transactionhead (transactionid, transactionno, operatorid, transdate) values (2, '0120201202000001', 2, '2020-12-11 00:00:00');
insert into demo.transactionhead (transactionid, transactionno, operatorid, transdate) values (3, '0120201202000002', 2, '2020-12-12 00:00:00');
select * from demo.transactionhead;
当前数据：
mysql> SELECT * FROM demo.transactionhead;
+---------------+------------------+------------+---------------------+
| transactionid | transactionno | operatorid | transdate |
+---------------+------------------+------------+---------------------+
| 1 | 0120201201000001 | 1 | 2020-12-10 00:00:00 |
| 2 | 0120201202000001 | 2 | 2020-12-11 00:00:00 |
| 3 | 0120201202000002 | 2 | 2020-12-12 00:00:00 |
+---------------+------------------+------------+---------------------+
mysql> SELECT * FROM demo.transactiondetails;
+---------------+------------+----------+-------+------------+
| transactionid | itemnumber | quantity | price | salesvalue |
+---------------+------------+----------+-------+------------+
| 1 | 1 | 1.000 | 89.00 | 89.00 |
| 1 | 2 | 2.000 | 5.00 | 10.00 |
| 2 | 1 | 2.000 | 89.00 | 178.00 |
| 3 | 2 | 10.000 | 5.00 | 50.00 |
+---------------+------------+----------+-------+------------+
mysql> SELECT * FROM demo.operator;
+------------+----------+--------+--------------+-------------+---------+--------------------+--------+
| operatorid | branchid | workno | operatorname | phone | address | pid | duty |
+------------+----------+--------+--------------+-------------+---------+--------------------+--------+
| 1 | 1 | 001 | 张静 | 18612345678 | 北京 | 110392197501012332 | 店长 |
| 2 | 1 | 002 | 李强 | 13312345678 | 北京 | 110222199501012332 | 收银员 |
+------------+----------+--------+--------------+-------------+---------+--------------------+--------+
mysql> SELECT * FROM demo.goodsmaster;
+------------+---------+-----------+---------------+------+-----------+
| itemnumber | barcode | goodsname | specification | unit | saleprice |
+------------+---------+-----------+---------------+------+-----------+
| 1 | 0001 | 书 | | 本 | 89.00 |
| 2 | 0002 | 笔 | | 支 | 5.00 |
+------------+---------+-----------+---------------+------+-----------+
查询：
mysql> SELECT a.transdate, c.operatorname, d.goodsname, b.quantity, b.price, b.salesvalue
-> FROM demo.transactionhead AS a
-> JOIN demo.transactiondetails AS b ON (a.transactionid = b.transactionid)
-> JOIN demo.operator AS c ON (a.operatorid = c.operatorid)
-> JOIN demo.goodsmaster AS d ON (b.itemnumber = d.itemnumber);
+---------------------+--------------+-----------+----------+-------+------------+
| transdate | operatorname | goodsname | quantity | price | salesvalue |
+---------------------+--------------+-----------+----------+-------+------------+
| 2020-12-10 00:00:00 | 张静 | 书 | 1.000 | 89.00 | 89.00 |
| 2020-12-10 00:00:00 | 张静 | 笔 | 2.000 | 5.00 | 10.00 |
| 2020-12-11 00:00:00 | 李强 | 书 | 2.000 | 89.00 | 178.00 |
| 2020-12-12 00:00:00 | 李强 | 笔 | 10.000 | 5.00 | 50.00 |
+---------------------+--------------+-----------+----------+-------+------------+
如果想看看每天的销售数量和销售金额，可以按照一个字段 transdate 对数据进行分组和统计：
mysql> select a.transdate, sum(b.quantity), sum(b.salesvalue)
-> from demo.transactionhead as a
-> join demo.transactiondetails as b on (a.transactionid = b.transactionid)
-> group by a.transdate;
+---------------------+-----------------+-------------------+
| transdate | sum(b.quantity) | sum(b.salesvalue) |
+---------------------+-----------------+-------------------+
| 2020-12-10 00:00:00 | 3.000 | 99.00 |
| 2020-12-11 00:00:00 | 2.000 | 178.00 |
| 2020-12-12 00:00:00 | 10.000 | 50.00 |
+---------------------+-----------------+-------------------+
如果想看每天、每个收银员的销售数量和销售金额，就可以按 2 个字段进行分组和统计，分别是 transdate 和 operatorname：
mysql> select a.transdate, c.operatorname, sum(b.quantity), sum(b.salesvalue)
-> from demo.transactionhead as a
-> join demo.transactiondetails as b on (a.transactionid = b.transactionid)
-> join demo.operator as c on (a.operatorid = c.operatorid)
-> group by a.transdate, c.operatorid;
+---------------------+--------------+-----------------+-------------------+
| transdate | operatorname | sum(b.quantity) | sum(b.salesvalue) |
+---------------------+--------------+-----------------+-------------------+
| 2020-12-10 00:00:00 | 张静 | 3.000 | 99.00 |
| 2020-12-11 00:00:00 | 李强 | 2.000 | 178.00 |
| 2020-12-12 00:00:00 | 李强 | 10.000 | 50.00 |
+---------------------+--------------+-----------------+-------------------+
通过对销售数据按照交易日期和收银员进行分组，再对组内数据进行求和统计，就实现了对每天、每个收银员的销售数量和销售金额的查询。
having
回到开头的超市经营者的需求：查询单笔销售金额超过 50 元的商品：
mysql> select a.goodsname
-> from demo.goodsmaster as a
-> join demo.transactiondetails as b on (a.itemnumber = b.itemnumber)
-> group by a.goodsname
-> having max(b.salesvalue) > 50;
+-----------+
| goodsname |
+-----------+
| 书 |
+-----------+
1 row in set (0.00 sec)
这种查询方式在 MySQL 里面是分四步实现的：
第一步，把流水明细表和商品信息表通过公共字段 itemnumber 连接起来，从 2 个表中获取数据：
mysql> select a.*, b.*
-> from demo.transactiondetails as a
-> join demo.goodsmaster as b on (a.itemnumber = b.itemnumber);
+---------------+------------+----------+-------+------------+------------+---------+-----------+---------------+------+-----------+
| transactionid | itemnumber | quantity | price | salesvalue | itemnumber | barcode | goodsname | specification | unit | saleprice |
+---------------+------------+----------+-------+------------+------------+---------+-----------+---------------+------+-----------+
| 1 | 1 | 1.000 | 89.00 | 89.00 | 1 | 0001 | 书 | | 本 | 89.00 |
| 1 | 2 | 2.000 | 5.00 | 10.00 | 2 | 0002 | 笔 | | 支 | 5.00 |
| 2 | 1 | 2.000 | 89.00 | 178.00 | 1 | 0001 | 书 | | 本 | 89.00 |
| 3 | 2 | 10.000 | 5.00 | 50.00 | 2 | 0002 | 笔 | | 支 | 5.00 |
+---------------+------------+----------+-------+------------+------------+---------+-----------+---------------+------+-----------+
第二步，把结果集按照商品名称（itemnumber）分组：
组 1： 
组 2：
第三步，对分组后的数据集进行筛选，把组中字段 salesvalue 的最大值 >50 的组筛选出来：
第四步，返回商品名称。这时就得到了需要的结果：单笔销售金额超过 50 元的商品就是“书”。
简单小结下使用 HAVING 的查询过程。
- 首先，把所有的信息都准备好，包括从关联表中获取需要的信息，对数据集进行分组，形成一个包含所有需要的信息的数据集合
- 接着，再通过 HAVING 条件的筛选，得到需要的数据
正确使用where和having
两者的区别：
-
如果需要通过连接从关联表中获取需要的数据，where是先筛选后连接，而having是先连接后筛选
这一点，就决定了在关联查询中，
where比having更高效。因为where可以先筛选，用一个筛选后的较小数据集和关联表进行连接，这样占用的资源比较少，执行效率也就比较高。having则需要先把结果集准备好，也就是用未被筛选的数据集进行关联，然后对这个大的数据集进行筛选，这样占用的资源就比较多，执行效率也较低。 -
where可以直接使用表中的字段作为筛选条件，但不能使用分组中的计算函数作为筛选条件；having必须要与group by配合使用，可以把分组计算的函数和分组字段作为筛选条件这决定了，在需要对数据进行分组统计的时候，
having可以完成where不能完成的任务。这是因为，在查询语法结构中，where在group by之前，所以无法对分组结果进行筛选。having在group by之后，可以使用分组字段和分组中的计算函数，对分组的结果集进行筛选，这个功能是where无法完成的。
假如超市经营者提出，要查询一下是哪个收银员、在哪天卖了 2 单商品。这种必须先分组才能筛选的查询，用 where 语句实现就比较难，我们可能要分好几步，通过把中间结果存储起来，才能搞定。但是用 having，则很轻松，代码如下：
mysql> select a.transdate, c.operatorname
-> from demo.transactionhead as a
-> join demo.transactiondetails as b on (a.transactionid = b.transactionid)
-> join demo.operator as c on (a.operatorid = c.operatorid)
-> group by a.transdate, c.operatorid
-> having count(*) = 2;
+---------------------+--------------+
| transdate | operatorname |
+---------------------+--------------+
| 2020-12-10 00:00:00 | 张静 |
+---------------------+--------------+
where 和 having 各自的优缺点：
| 优点 | 缺点 | |
|---|---|---|
| where | 先筛选数据，再关联，执行效率高 | 不能使用分组中俄计算函数进行筛选 |
| having | 可以使用分组中的计算函数 | 在最后的结果集中进行筛选，执行效率较低 |
需要注意的是，where 和 having 也不是互相排斥的，我们可以在一个查询里面同时使用where 和 having。
举个例子，假设现在有一组销售数据，包括交易时间、收银员、商品名称、销售数量、价格和销售金额等信息，超市的经营者要查询“2020-12-10”和“2020-12-11”这两天收银金额超过 100 元的销售日期、收银员名称、销售数量和销售金额。
mysql> select a.transdate, c.operatorname, d.goodsname, b.quantity, b.price, b.salesvalue
-> from demo.transactionhead as a
-> join demo.transactiondetails as b on (b.transactionid = a.transactionid)
-> join demo.operator as c on (c.operatorid = a.operatorid)
-> join demo.goodsmaster as d on (d.itemnumber = b.itemnumber);
+---------------------+--------------+-----------+----------+-------+------------+
| transdate | operatorname | goodsname | quantity | price | salesvalue |
+---------------------+--------------+-----------+----------+-------+------------+
| 2020-12-10 00:00:00 | 张静 | 书 | 1.000 | 89.00 | 89.00 |
| 2020-12-10 00:00:00 | 张静 | 笔 | 2.000 | 5.00 | 10.00 |
| 2020-12-11 00:00:00 | 李强 | 书 | 2.000 | 89.00 | 178.00 |
| 2020-12-12 00:00:00 | 李强 | 笔 | 10.000 | 5.00 | 50.00 |
+---------------------+--------------+-----------+----------+-------+------------+
分析需求：由于是要按照销售日期和收银员进行统计，所以，必须按照销售日期和收银员进行分组，因此，我们可以通过使用group by和having进行查询：
mysql> select a.transdate, c.operatorname, sum(b.quantity), sum(b.salesvalue)
-> from demo.transactionhead as a
-> join demo.transactiondetails as b on (b.transactionid = a.transactionid)
-> join demo.operator as c on (c.operatorid = a.operatorid)
-> group by a.transdate, c.operatorname
-> having a.transdate in ('2020-12-10' , '2020-12-11') and sum(b.salesvalue) > 100;
+---------------------+--------------+-----------------+-------------------+
| transdate | operatorname | sum(b.quantity) | sum(b.salesvalue) |
+---------------------+--------------+-----------------+-------------------+
| 2020-12-11 00:00:00 | 李强 | 2.000 | 178.00 |
+---------------------+--------------+-----------------+-------------------+
其实having后面的筛选条件，就会发现，条件 a.transdate IN ('2020-12-10' , '2020-12-11')，其实可以用where来限定：
mysql> select a.transdate, c.operatorname, sum(b.quantity), sum(b.salesvalue)
-> from demo.transactionhead as a
-> join demo.transactiondetails as b on (b.transactionid = a.transactionid)
-> join demo.operator as c on (c.operatorid = a.operatorid)
-> where a.transdate in ('2020-12-10' , '2020-12-11') -- 先按日期筛选
-> group by a.transdate, c.operatorname
-> having sum(b.salesvalue) > 100; -- 后按金额筛选
+---------------------+--------------+-----------------+-------------------+
| transdate | operatorname | sum(b.quantity) | sum(b.salesvalue) |
+---------------------+--------------+-----------------+-------------------+
| 2020-12-11 00:00:00 | 李强 | 2.000 | 178.00 |
+---------------------+--------------+-----------------+-------------------+
得到了需要的结果，这是因为把条件拆分开，包含分组统计函数的条件用having，普通条件用where。这就既利用了where条件的高效快速，又发挥了having可以使用包含分组统计函数的查询条件的优点。当数据量特别大的时候，运行效率会有很大的差别。