PostgreSQL 一行变多行详解数据库

实例表： 

postgres=# /d book 
                              Table "public.book" 
 Column |       Type        |                     Modifiers                      
--------+-------------------+--------------------------------------------------- 
 id     | integer           | not null default nextval('book_id_seq'::regclass) 
 name   | character varying |  
 tag    | character varying | 

postgres=# select * from book; 
 id | name |   tag     
----+------+---------- 
  1 | java | aa,bb,cc 
  2 | C++  | dd,ee 
(2 rows)

需要的效果： 

 name | unnest  
------+-------- 
 java | aa 
 java | bb 
 java | cc 
 C++  | dd 
 C++  | ee 
(5 rows)

方法1： unnest(数组，分隔符)

postgres=# select name,unnest(string_to_array(tag,',')) from book; 
 name | unnest  
------+-------- 
 java | aa 
 java | bb 
 java | cc 
 C++  | dd 
 C++  | ee 
(5 rows)

postgres=# select string_to_array(tag,',') from book; 
 string_to_array  
----------------- 
 {aa,bb,cc} 
 {dd,ee} 
(2 rows)

方法2：regexp_split_to_table(字符串，分隔符)

postgres=# select name,regexp_split_to_table(tag,',') from book; 
 name | regexp_split_to_table  
------+----------------------- 
 java | aa 
 java | bb 
 java | cc 
 C++  | dd 
 C++  | ee 
(5 rows)

另外，字符串变数组的方法：

postgres=# select name,string_to_array(tag,',') from book; 
 name | string_to_array  
------+----------------- 
 java | {aa,bb,cc} 
 C++  | {dd,ee} 
(2 rows)

postgres=# select name,regexp_split_to_array(tag,',') from book; 
 name | regexp_split_to_array  
------+----------------------- 
 java | {aa,bb,cc} 
 C++  | {dd,ee} 
(2 rows)