题目：
CREATE TABLE plch_departments (
department_id INTEGER PRIMARY KEY
,department_name VARCHAR2 (100))
/
BEGIN
INSERT INTO plch_departments
VALUES (100, 'Marketing');
INSERT INTO plch_departments
VALUES (200, 'Catering');
COMMIT;
END;
/
CREATE TABLE plch_employees
(
employee_id INTEGER
, last_name VARCHAR2 (100)
, salary NUMBER
, department_id INTEGER REFERENCES plch_departments (department_id)
)
/
BEGIN
INSERT INTO plch_employees
VALUES (100
, 'Jobs'
, 1000000
, 100);
INSERT INTO plch_employees
VALUES (200
, 'Ellison'
, 1000000
, 200);
INSERT INTO plch_employees
VALUES (300
, 'Gates'
, 1000000
, 200);
COMMIT;
END;
/
Which of the choices contain a query that returns the same dataset as the following query:
SELECT e.last_name, e.salary, d.department_name
FROM plch_employees e, plch_departments d
WHERE e.department_id = d.department_id
大致解释就是说：
有两个表，plch_departments 和 plch_employees ，共有一个相同字段 department_id ；
现在要将两个表格中三个字段一起查询输出，题目中提供的方法是 ：
通过 WHERE 条件语句进行比较两表中的department_id，相同时则输出对应的字段。
Choice 1
SELECT e.last_name, e.salary, d.department_name
FROM plch_employees e NATURAL JOIN plch_departments d
Use the Natural Join syntax to join two tables based on column names they have in common, in this case department_id.
即 NATURAL JOIN 时两表必须有一样的列
Choice 2
SELECT e.last_name, e.salary, d.department_name
FROM plch_employees e JOIN plch_departments d USING (department_id)
JOIN…USING is much like the Natural Join, but in this case you specify the columns to use in the join.
USING 后的字段即公共字段
Choice 3
选错了，字段必须用括号括起来，我以为有不有都可以
Choice 4
SELECT e.last_name, e.salary, d.department_name
FROM plch_employees e
JOIN
plch_departments d
ON (e.department_id = d.department_id)
The JOIN…ON syntax is used to join table on the fields you specify. They need not be of the same name and as long as there can be an implicit typecast, they don’t even have to be of the same type.
JOIN....ON不限制字段名和字段类型必须相同
Choice 5
SELECT e.last_name, e.salary, d.department_name
FROM plch_employees e CROSS JOIN plch_departments d
A Cross Join joins every record in the first table with every record in the second table therefore resulting in a Cartesian product, which does not give us the result we wanted for this quiz.
Of course, there may come a time when you actually do need a Cartesian product. In this case, the ANSI syntax is more explicit and manageable than the ‘old’ Oracle syntax - with which Cartesian products are the result of leaving off predicates in the WHERE clause.
CROSS JOIN 将第一个表中的每个记录与第二个表中的每个记录联接在一起，产生笛卡尔积。（借助Google翻译的）
这里不知道笛卡尔积，可以看下下图
Choice 6
这里也选错了，看到cross join 就没选了。其实应该看where条件语句，限制了笛卡尔积的范围，毕竟这里的字段department_id值对应是一对一的，一个 100 200，一个 100 200 300，cross join没有啥影响。
SELECT e.last_name, e.salary, d.department_name
FROM plch_employees e CROSS JOIN plch_departments d
WHERE e.department_id = d.department_id
A Cross Join joins every record in the first table with every record in the second table. The where clause then limits the result of the query. This choice looks an awful lot like the ‘old’ Oracle syntax, just a bit more verbose.
De advised though that creating a cartesian product and then filtering it may not be very performant and may put a big stress on the available memory.
不过这里也说了下，比较冗余的查询语句，造成内存负担。估计除了出题时会这么写，实际不会这么操作吧。
扩展阅读Oracle9i表连接
