数据库关系代数
基本概念
元祖行：水平方向上每一行为一条记录，这个记录对应1个实体。一般称为元祖，元祖行或者记录
属性列：垂直方向上每一列为一个属性，一般称为属性列，字段等。关系表达式中可以用列序号代替列名
表结构：属性列个数与排列，属性列个数也称为表的目数。
并交叉都是二元运算符
并（∪）：综合2个表的所有元组，且去除重复。
交（∩）：列出2个表都有的元祖
差（-）：以左侧对象为主，在左侧对象的元祖集合中去除右侧对象已有的元祖，则为差
| 关系s1 | 关系s2 | |||||
| sno | sname | sdept | sno | sname | sdept | |
| no0001 | mary | is | no0001 | mary | is | |
| n00003 | candy | is | no0008 | katter | is | |
| no0004 | jam | is | no0021 | tom | is | |
s1并s2的结果为
| s1∪s2 | ||
| sno | sname | sdept |
| no0001 | mary | is |
| n00003 | candy | is |
| no0004 | jam | is |
| no0008 | katter | is |
| no0021 | tom | is |
s1∩s2
| s1∩s2 | ||
| sno | sname | sdept |
| no0001 | mary | is |
s1-s2
| s1-s2 | ||
| sno | sname | sdept |
| n00003 | candy | is |
| no0004 | jam | is |
笛卡尔积（）：二元运算，属性列是二者之和，元祖行数是二者乘积。元祖结果是以左侧对象的元祖分别拼接右侧对象的元祖，形成不同的排列组合。不同表的同名属性，需要以“表名.列名”加以区分。
| s1×s2 | |||||
| s1.sno | s1.sname | s1.sdept | s2.sno | s2.sname | s2.sdept |
| no0001 | mary | is | no0001 | mary | is |
| no0001 | mary | is | no0008 | katter | is |
| no0001 | mary | is | no0021 | tom | is |
| n00003 | candy | is | no0001 | mary | is |
| n00003 | candy | is | no0008 | katter | is |
| n00003 | candy | is | no0021 | tom | is |
| no0004 | jam | is | no0001 | mary | is |
| no0004 | jam | is | no0008 | katter | is |
| no0004 | jam | is | no0021 | tom | is |
1650
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
