我在商店里有500个售货亭,并且有一个表根据商店ID跟踪每个打印的时间戳。我需要生成一个记录集,给出打印之间的时间间隔(以秒为单位)。所有这些数据都存储在一个表中。为每个打印插入一条记录,其中包含商店ID和timeStamp。
表名=打印
id store_id timestamp
2013-3-1 00:01:01:00:01:00:01
2013-3-1 00:01**2: 00:00:01 **
2013-3-1 00:01:00:01
2013-3-1 00:12:4-2:00:00-12**
2013-3-1 00:06:00:06
2013-3-1 00:15:6-2:00:00-11:00**
我需要拉取所有商店2号指纹之间的时间。*是为了便于您找到我需要比较的记录。
RecordSet结果如下:
id store_id myTimeDiffSeconds
*0
4/2* 11
6/2*3
这需要简单和快速。我可以在没有临时表的情况下做到这一点吗?
发布于 2013-03-27 18:25:01
您可以通过两种方式编写查询
第一个正是ethrbunny指出的
mysql> SELECT t1.id,
t1.store_id,
timestampdiff(second,IFNULL( (SELECT MAX(t2.timestamp)
FROM print t2
WHERE t2.store_id=2
AND t2.timestamp< t1.timestamp)
,t1.timestamp),t1.timestamp) myTimeDiffSeconds
FROM print t1
WHERE t1.store_id=2 ORDER BY t1.timestamp;
+------+----------+-------------------+
| id | store_id | myTimeDiffSeconds |
+------+----------+-------------------+
| 2 | 2 | 0 |
| 4 | 2 | 11 |
| 6 | 2 | 3 |
+------+----------+-------------------+
3 rows in set (0.00 sec)
另一种方法是使用会话变量来保存上一次时间,但在这种情况下,我们需要第一次获得最小时间戳
mysql> select min(p.timestamp) into @prev_time from print p where p.store_id=2;
Query OK, 1 row affected (0.00 sec)
mysql> select id,
store_id,
timestampdiff(second,@prev_time,timestamp) myTimeDiffSeconds,
@prev_time:=timestamp
from print
where store_id=2 order by timestamp;
+------+----------+-------------------+---------------------+
| id | store_id | myTimeDiffSeconds | @prev_time:=t |
+------+----------+-------------------+---------------------+
| 2 | 2 | 0 | 2013-03-01 00:00:01 |
| 4 | 2 | 11 | 2013-03-01 00:00:12 |
| 6 | 2 | 3 | 2013-03-01 00:00:15 |
+------+----------+-------------------+---------------------+
3 rows in set (0.00 sec)
index on (timestamp,store_id)将使查询执行得更好。
https://stackoverflow.com/questions/15620476
复制相似问题