0. Contents summary ：

        a. The concept of join query , classification

        b. Connection query principle , The cartesian product （ Two tables ）

        c. Two table join query syntax , Equivalent connection , Non equivalent connection , Self join

        d. External connection ： Left outer connection and right outer connection ; The relationship between outer connection and inner connection

        e. Multiple table joins

1. Join query concepts and classifications ：

Link query , Or cross table query , Different from the previous single table query , It refers to the combination of multiple tables to query data . According to the grammatical age , It is divided into SQL1992 and SQL1999; According to the connection object , It is divided into internal connection query , External connection query and full connection query （ Leave out ）. Internal connection query and External connection query The concept of is put in the external connection query to solve .

2. Connect query principle and Cartesian product phenomenon

Cartesian phenomena occur in unconditional join queries , It embodies the essence of join query . Let's take the internal connection between two tables as an example .

Have a watch emp：

  surface dep：

  Make internal connections ：

select e.deptno, d.dname from emp as e, dep as d;

The final result is 56 Table of rows , The table is too long to show . Look at the table and you can find ,emp Every record in the table is related to dep All in the table 4 Records match the past one by one , This led to the final number of records 14 * 4 = 56 strip . We call this phenomenon Cartesian product phenomenon , That is, the number of records of the table connection result is equal to the product of the number of records of each table .

 3. Two table join query syntax

In order to further improve the efficiency of the third statement in the table , We can table1,table2 names , And used in field1 and field2 Before ：

select t1.field1,t2.field2 from table1 as t1 inner join table2 as t2 on condition;

stay field1 and field2 Specifying the table name prevents mysql At the same time t1 and t2 Search for field1（field2）, So it improves efficiency . Otherwise ,mysql Will be in t1 and t2 Search for field1, Will also be in t1 and t2 Search for field2, This reduces efficiency . in addition , added condition after , Although from the results , We avoid the Cartesian product phenomenon . however , The number of matches between the records of the two tables has not decreased , It is still the product of the number of records in each table .

We can think of it like this ： The above statement is executed first from and join sentence , Let the two tables be connected into a new table according to the Cartesian product phenomenon , And then execute on sentence , Meet in screening condition Field of . The last is select.        therefore , Cartesian product embodies table connection （ Take inner connection as an example ） Principle , Very important .

Now consider condition 了 . For internal connections , If condition Is an equality condition , Then call this internal connection Equivalent connection . for example ： Find the working place of each employee .

select e.ename, d.dname from emp as e join dep as d on e.deptno = d.deptno;

The condition here is ：e.deptno = d.deptno, It's an equation . So this inner connection is an equivalent connection .

In use enterprise Import another table into the database salgrade：

  We unite emp and salgrade Find employee salary grade . This requires internal connections Non equivalent connection （condition Is a non equality condition ）.

select e.ename, s.grade 
from emp as e 
join salgrade as s 
on e.sal between s.losal and s.hisal;

Self join It is a table connected with itself to query data . The idea is to treat one table as two tables . The syntax is consistent with equivalent connection and non equivalent connection , All conform to the syntax of inner connection . for example ： Inquire about the leader of the employee .

select e1.ename, e2.ename 
from emp as e1 
inner join emp as e2 
on e1.mgr = e2.empno;

4. Two off meter connections

The difference between two tables' external connection and internal connection lies in whether there is a master-slave relationship between two tables . For internal connections , The two tables have the same status ,select After that, only （ Cartesian product ） Qualified records ; For external connections , There is a master-slave relationship between the two tables . The external connection mainly checks the main table , All records of the main table will be displayed , Whether qualified or not ; It will display the qualified records in the table . The grammar is ：

// Left connection 
select t1.field1, t2.field2 from table1 t1 left join table2 t2 on condition;
// The right connection 
select t1.field1, t2.field2 from table1 t1 right join table2 t2 on condition;

Left link selected join The table on the left is the main table , On the right is the slave table ; The right connection is the opposite . Left and right connections must be able to transform each other .

Example ： Query the leader names of all employees , With or without leadership .

because King There is no leadership , The result should have been 13 Bar record ; But after using the syntax of outer connection , Yes 14 Bar record .

mysql> select e1.ename,e2.ename
    -> from emp e1
    -> left outer join emp e2
    -> on e1.mgr = e2.empno;

  Do not conform to the on Later conditional King This record also exists , This is the difference between external connection and internal connection ： It displays all records of the master table and qualified slave table records . The outer connection can be considered as the inner connection based on the main table .

5. Multi-table query （ Three watches or above ）

grammar ：

select
    ......
from 
    table1 t1
(inner/left/right)join
    table2 t2
on
    (t1,t2)condition
(inner/left/right)join
    table3 t3
on
    (t1,t2,t3)condition
...
...
// The first question is , Every time the condition can only be join The following table name and from Is it composed of the following table name ？ This time join Table name after 
// And all table names that have appeared before ？
// The experimental results are all acceptable .join tablen tn on (t1,t2,t3,...,tn)

In multiple tables , Internal and external connections can be mixed . But here's the thing ,left The connection does not just take the previous table as the main table , Empathy right The connection does not just take the following table as the main table .

Select the employee name , Corresponding department name , Salary and salary grade . It is required to print out all departments .

select e.ename, d.dname, e.sal, s.grade
    from emp e
    right outer join dep d
    on e.deptno = d.deptno
    join salgrade s
    on e.sal between s.losal and s.hisal;
//e With the first d External connection ,d Main table ; The result is due to s Internal connection , The front outer connection was screened out . So the final result is not what we want .

The result is ：

Incompatible with the purpose .

  Only in this way can we achieve our goal ：

mysql> select e.ename, d.dname, e.sal, s.grade
    -> from emp e
    -> join salgrade s
    -> on e.sal between s.losal and s.hisal
    -> right outer join dep d
    -> on e.deptno = d.deptno
    -> \g
 or 
select e.ename, d.dname, e.sal, s.grade
    from emp e
    right outer join dep d
    on e.deptno = d.deptno
    left join salgrade s
    on e.sal between s.losal and s.hisal;

  I think the reasonable explanation is ：（ Suppose the table connected with this is table1,table2,...,tablen） First ,table1 Hui He table2 Connect (join), After performing on Screening . The results obtained after screening are compared with table3 Conduct join, We'll do it later on Later conditions . And so on . Of all the join and on After the implementation , perform from.

The rationality of this explanation lies in ：1. Conform to language intuition ;2. Connecting after filtering can improve execution efficiency ;3. Can explain the above phenomenon . This also tells us how to mix inner join and outer join in multi table query . If you want to use a table as the main table , Or put it after the beginning left join Or put it at the end right join, Or use it before it right join Use... After it left join.