postgres=# \d book2
                              Table "public.book2"
 Column |       Type        |                     Modifiers                      
--------+-------------------+----------------------------------------------------
 id     | integer           | not null default nextval('book2_id_seq'::regclass)
 name   | character varying | 
 tag    | character varying |

postgres=# select * from book2;
 id | name | tag 
----+------+-----
  1 | java | aa
  1 | java | bb
  1 | java | cc
  2 | C++  | dd
  2 | C++  | ee
(5 rows)

需要的效果：

 name | string_agg 
------+------------
 C++  | dd,ee
 java | aa,bb,cc
(2 rows)

方法1：string_agg(字段名，分隔符)

postgres=# select name,string_agg(tag,',') from book2 group by name;
 name | string_agg 
------+------------
 C++  | dd,ee
 java | aa,bb,cc
(2 rows)

方法2：array_agg(字段名)

postgres=# select name,array_agg(tag) from tb3 group by name;
 name | array_agg  
------+------------
 c++  | {dd,ee}
 java | {aa,bb,cc}
(2 rows)<strong>
</strong>postgres=# select name,array_to_string(array_agg(tag),',') from tb3 group by name;
 name | array_to_string 
------+-----------------
 c++  | dd,ee
 java | aa,bb,cc
(2 rows)

array_agg(distinct(字段名)) 拼接唯一的字段

实例表：postgres=# \d book2 Table "public.book2" Column | Type | Modifiers --------+-------------------+------------ alter table goods add primary key(sid); 2. 添加外键 alter table orders add foreign key(goods_id) references goods(sid) on update cascadeon delete cascade; on update cascade: 被引用行更新时，引用行自动更