SQL

Question

我有以下 table:

CREATE temp TABLE "t_table" (
  usr_id bigint,
  address varchar[],
  msg_cnt bigint,
  usr_cnt bigint,
  source varchar[],
  last_update timestamp
);

添加数据：

INSERT INTO "t_table"(usr_id, address, msg_cnt, usr_cnt, source, last_update) VALUES (1, '{44.154.48.125,81.134.82.111,95.155.38.120,94.134.88.136}', 10, 3, '{src1,src2}', '2019-10-16 22:16:22.163000');
INSERT INTO "t_table"(usr_id, address, msg_cnt, usr_cnt, source, last_update) VALUES (2, '{44.154.48.125}', 10, 3, '{src1,src3}', '2019-10-16 22:16:22.163000');
INSERT INTO "t_table"(usr_id, address, msg_cnt, usr_cnt, source, last_update) VALUES (3, '{94.134.88.136}', 10, 3, '{src1,src4}', '2019-10-16 22:16:22.163000');
INSERT INTO "t_table"(usr_id, address, msg_cnt, usr_cnt, source, last_update) VALUES (4, '{127.0.0.1}', 10, 3, '{src1,src5}', '2019-10-16 22:16:22.163000');
INSERT INTO "t_table"(usr_id, address, msg_cnt, usr_cnt, source, last_update) VALUES (5, '{127.0.0.1,5.5.5.5}', 10, 3, '{src1,src3}', '2019-10-16 22:16:22.163000');
INSERT INTO "t_table"(usr_id, address, msg_cnt, usr_cnt, source, last_update) VALUES (6, '{1.1.0.9}', 10, 3, '{src1,src2}', '2019-10-16 22:16:22.163000');

查找共享地址的用户。

预期结果：

|      users                      |  address                                                    | sum_msg_cnt  |  sum_usr_cnt     | max_last_date                  | source                      |
|---------------------------------|-------------------------------------------------------------|--------------|------------------|--------------------------------|-----------------------------|
|    {1,2,3}                      |  {44.154.48.125,81.134.82.111,95.155.38.120,94.134.88.136}  | 30           |         9        |  "2019-10-16 22:16:22.163000"  |    {src4,src1,src2,src3}    |
|    {4,5}                        |  {127.0.0.1,5.5.5.5}                                        | 20           |         6        |  "2019-10-16 22:16:22.163000"  |    {src1,src5,src3}         |
|    {6}                          |  {1.1.0.9}                                                  | 10           |         3        |  "2019-10-16 22:16:22.163000"  |    {src1,src2}              |

问题：

如何制定 SQL 查询以获得预期结果？

非常感谢。

更多信息：

PostgreSQL 9.5.19

Answer 1

使用 "group by" 运算符

对地址进行分组

Answer 2

我不知道这是否是最有效的方法，但我现在想不出更好的方法。

我认为这在更大的 table 上会有糟糕的表现。

with userlist as (
  select array_agg(t.usr_id) as users, 
         a.address
  from t_table t
    left join unnest(t.address) as a(address) on true
  group by a.address  
), shared_users as (
  select u.address,
         array(select distinct ul.uid
               from userlist u2, unnest(u2.users) as ul(uid)
               where u.users && u2.users
               order by ul.uid) as users
  from userlist u
)
select users, array_agg(distinct address)
from shared_users
group by users;

它有什么作用？

第一个 CTE 收集共享至少一个地址的所有用户。 userlist CTE 的输出是：

users | address      
------+--------------
{1}   | 95.155.38.120
{1,3} | 94.134.88.136
{1,2} | 44.154.48.125
{6}   | 1.1.0.9      
{4,5} | 127.0.0.1    
{1}   | 81.134.82.111
{5}   | 5.5.5.5

现在这可用于聚合那些共享至少一个地址的用户列表。 shared_users CTE 的输出是：

address       | users  
--------------+--------
95.155.38.120 | {1,2,3}
94.134.88.136 | {1,2,3}
44.154.48.125 | {1,2,3}
1.1.0.9       | {6}    
127.0.0.1     | {4,5}  
81.134.82.111 | {1,2,3}
5.5.5.5       | {4,5}

如您所见，我们现在有具有相同 usr_ids 列表的组。在最后一步中，我们可以按这些分组并聚合地址，然后 return:

users   | array_agg                                                
--------+----------------------------------------------------------
{1,2,3} | {44.154.48.125,81.134.82.111,94.134.88.136,95.155.38.120}
{4,5}   | {127.0.0.1,5.5.5.5}                                      
{6}     | {1.1.0.9}

Online example

SQL - 创建 SQL 加入列表

SQL - create SQL to join lists

postgresql

postgresql-9.5