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)

 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,cc}
 {dd,ee}
(2 rows)

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,cc}
C++ | {dd,ee}
(2 rows)

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