获取前行标识，而在ID

Question

排序没有提到下面如何从表以前&下一个ID具有结构：

+----+---------+---------------------+ 
| id | urgency | timestamp   | 
+----+---------+---------------------+ 
| 1 | 0  | 2013-01-01 00:00:00 | 
| 2 | 2  | 2013-01-01 00:00:00 | 
| 3 | 1  | 2013-01-05 09:30:00 | 
| 4 | 2  | 2013-01-01 00:00:00 | 
| 5 | 2  | 2013-01-01 00:00:00 | 
| 6 | 1  | 2013-01-06 10:00:00 | 
| 7 | 0  | 2013-01-01 00:00:00 | 
| 8 | 0  | 2013-01-03 00:00:00 | 
| 9 | 1  | 2013-02-01 13:30:00 | 
| 10 | 0  | 2013-01-04 00:00:00 | 
+----+---------+---------------------+ 

表进行排序如下：尿急递增，时间戳递增， ID ASC

这里是排序表：

+----+---------+---------------------+ 
| id | urgency | timestamp   | 
+----+---------+---------------------+ 
| 1 | 0  | 2013-01-01 00:00:00 | 
| 7 | 0  | 2013-01-01 00:00:00 | 
| 8 | 0  | 2013-01-03 00:00:00 | 
| 10 | 0  | 2013-01-04 00:00:00 | 
| 3 | 1  | 2013-01-05 09:30:00 | 
| 6 | 1  | 2013-01-06 10:00:00 | 
| 9 | 1  | 2013-02-01 13:30:00 | <= CURRENT_ID 
| 2 | 2  | 2013-01-01 00:00:00 | 
| 4 | 2  | 2013-01-01 00:00:00 | 
| 5 | 2  | 2013-01-01 00:00:00 | 
+----+---------+---------------------+ 

有没有办法从MySQL选择一些地方的位置CURRENT_ID - 1或CURRENT_ID + 1？

我能做的是执行一个简单的查询：

SELECT 
    id 
FROM 
    MY_TABLE 
ORDER BY 
    urgency asc, 
    timestamp asc, 
    id asc 

通过创建结果PHP一个循环，并寻找合适的位置，但是这是一个表，将会增加非常快。因此，这不是一个选项...希望有另一种解决方案，您可以建议

Answer 1

对于以前：

SELECT prev.id 
FROM my_table current 
JOIN my_table prev 
    ON (prev.urgency < current.urgency) 
    OR (prev.urgency = current.urgency 
     AND (prev.timestamp < current.timestamp 
      OR (prev.timestamp = current.timestamp AND prev.id < current.id))) 
WHERE current.id = @currentId 
ORDER BY prev.urgency desc, prev.timestamp desc, prev.id desc 
LIMIT 1 

对于未来：

SELECT next.id 
FROM my_table current 
JOIN my_table next 
    ON (next.urgency > current.urgency) 
    OR (next.urgency = current.urgency 
     AND (next.timestamp > current.timestamp 
      OR (next.timestamp = current.timestamp AND next.id > current.id))) 
WHERE current.id = @currentId 
ORDER BY next.urgency asc, next.timestamp asc, next.id asc 
LIMIT 1 

Sqlfiddle这里。

试图让它们在一个查询中更具挑战性，但可以使用类似的方法使用嵌套分组代替LIMIT来完成。我怀疑你会在这样一个复杂的查询中看到很多性能优势，你肯定会很难保持它...