可以为oracle数据库优化此sql查询吗？

Question

消息表

|-----|-------------|-----------------|--------|--------------|
|  id | message_id  | msg_connect_id  |  body  |  created_ts  |  
|-----|-------------|-----------------|--------|--------------|
|  1  | 12345       | 1234            |  body1 |  timestamp1  |
|  2  | 12346       | 1234            |  body2 |  timestamp2  |
|  3  | 12347       | 1234            |  body3 |  timestamp3  |
|  4  | 12348       | 5678            |  body4 |  timestamp4  |
|  5  | 12349       | 1234            |  body5 |  timestamp5  |
|-----|-------------|-----------------|--------|--------------|

是否提供了message_id作为查询参数，可以优化此查询吗？

SELECT id,message_id ,body,created_ts FROM message
where msg_connect_id = (select msg_connect_id from message where message_id ='12346')
and id <=  (select id from message where message_id ='12346') order by created_ts desc

我期望与message_id参数匹配的行以及它下面的所有具有相同msg_connect_id的行。 所以例如如果参数message_id为12346，则预期的行是ID为2、3和5的行。

Answer 1

您可以将子查询移至from子句并使用索引：

select m.id, m.message_id, m.body, m.created_ts 
from message m cross join
     (select m2.*
      from message m2
      where message_id = 12346  -- it looks like a number, so I assume it is
     ) mm
where m.msg_connect_id = mm.msg_connect_id and
      m.id <= mm.id
order by m.created_ts desc;

为此，您希望在(message_id)和(msg_connect_id, id)上建立索引。