MySQL Query to group on max event_count?

Question

I need to calculate the time difference between groups of events. Each "event group" starts over when the event_count column = 1, and the max event_count can be any number greater than 1.

How can I do this in MySQL?

Here is the table and some test data:

Table

CREATE TABLE `monitoring` ( 
  `event_id` int(10) unsigned NOT NULL AUTO_INCREMENT, 
  `hosting_id` char(36) DEFAULT NULL, 
  `event_timestamp` datetime DEFAULT NULL, 
  `event_type` tinyint(3) unsigned DEFAULT NULL, 
  `event_count` int(6) DEFAULT NULL, 
  `session` varchar(36) DEFAULT NULL, 
  PRIMARY KEY (`event_id`) 
) ENGINE=MyISAM AUTO_INCREMENT=1765 DEFAULT CHARSET=utf8;

Test Records:

INSERT INTO `monitoring` (`event_id`, `hosting_id`, `event_timestamp`, `event_type`, `event_count`, `session`)
VALUES
    (1753,'97948a60-2e44-d39c-bb02-506c80c97df4','2013-10-23 20:01:19',1,1,NULL),
    (1711,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 16:02:20',1,1,NULL),
    (1712,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 16:02:26',1,2,NULL),
    (1713,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 16:02:30',1,3,NULL),
    (1714,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 16:02:33',1,4,NULL),
    (1715,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 16:02:45',1,5,NULL),
    (1716,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 16:02:48',1,6,NULL),
    (1717,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 16:02:51',1,7,NULL),
    (1718,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 16:49:29',1,1,NULL),
    (1719,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 16:49:31',1,2,NULL),
    (1720,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 16:49:33',1,3,NULL),
    (1721,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 16:49:35',1,4,NULL),
    (1722,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 16:49:38',1,5,NULL),
    (1723,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 16:49:40',1,6,NULL),
    (1724,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 16:49:43',1,7,NULL),
    (1725,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 16:49:46',1,8,NULL),
    (1726,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 16:49:48',1,9,NULL),
    (1727,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 16:49:50',1,10,NULL),
    (1754,'988df609-4ae4-3062-6361-4fd2665cfb42','2013-10-24 02:48:32',1,1,NULL),
    (1729,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 17:02:28',1,1,NULL),
    (1730,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 17:03:25',1,2,NULL),
    (1731,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 17:03:30',1,3,NULL),
    (1732,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 17:05:42',1,4,NULL),
    (1733,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 17:06:03',1,5,NULL),
    (1758,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-24 12:23:43',1,1,NULL),
    (1757,'d425f99b-0812-5902-37da-4fd266c64545','2013-10-24 08:16:26',1,1,NULL),
    (1756,'20e00e08-459b-b229-d341-4fd266c23df1','2013-10-24 06:32:10',1,1,NULL),
    (1738,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 17:09:02',1,1,'5c31a1eb481bf9e0753f9a25538b0257'),
    (1739,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 17:09:04',1,2,'5c31a1eb481bf9e0753f9a25538b0257'),
    (1740,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 17:09:06',1,3,'5c31a1eb481bf9e0753f9a25538b0257'),
    (1741,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 17:09:07',1,4,'5c31a1eb481bf9e0753f9a25538b0257'),
    (1742,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 17:09:08',1,5,'5c31a1eb481bf9e0753f9a25538b0257'),
    (1743,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 17:09:09',1,6,'5c31a1eb481bf9e0753f9a25538b0257'),
    (1755,'988df609-4ae4-3062-6361-4fd2665cfb42','2013-10-24 06:30:20',1,1,NULL),
    (1745,'a9e17a2a-42d9-979f-0484-4fd2665b487a','2013-10-23 17:09:28',1,1,'5c31a1eb481bf9e0753f9a25538b0257'),
    (1746,'a9e17a2a-42d9-979f-0484-4fd2665b487a','2013-10-23 17:09:38',1,2,'5c31a1eb481bf9e0753f9a25538b0257'),
    (1747,'a9e17a2a-42d9-979f-0484-4fd2665b487a','2013-10-23 17:09:55',1,3,'5c31a1eb481bf9e0753f9a25538b0257'),
    (1748,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 17:13:54',1,1,NULL),
    (1749,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 17:13:56',1,2,NULL),
    (1750,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 17:13:58',1,3,NULL),
    (1751,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-23 17:14:00',1,4,NULL),
    (1752,'ecd5f06a-7b82-4cc8-deb2-4fd266eeec3e','2013-10-23 19:06:42',1,1,NULL),
    (1759,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-24 12:23:48',1,2,NULL),
    (1760,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-24 12:23:51',1,3,NULL),
    (1761,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-24 12:23:53',1,4,NULL),
    (1762,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-24 12:23:56',1,5,NULL),
    (1763,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-24 12:23:58',1,6,NULL),
    (1764,'c47204d5-d682-9ba4-b8b6-4fd26660651a','2013-10-24 12:24:23',1,7,NULL);

What I am aiming for is:

On 10-23-2013 down for 31 seconds 
On 10-23-2013 down for 21 seconds 
On 10-23-2013 down for 215 seconds
    etc...

which I will do once I can get this query figured out properly.

My Query so far is:

SELECT min(date(`event_timestamp`)) as EventGroupDate, (max(`event_timestamp`) - min(`event_timestamp`)) HowLong
FROM (
    SELECT `event_timestamp`, @eventGroup := @eventGroup + (`event_count` = 1) eventGroup
    FROM `monitoring`, (SELECT @eventGroup := 0) init
    WHERE `hosting_id` = 'c47204d5-d682-9ba4-b8b6-4fd26660651a'
    ORDER BY `event_timestamp`
) s
GROUP BY eventGroup

but wrong calculations. Exact calculations are below

and I will post updates to it as I make them.

Each event group starts as event_count = 1, and can end at any number greater than 1. Since this needs to sort by the event_timestamp it can be assumed that all in between could be considered a group. for example: record IDs 1711-1717 would be group 1, 1718-1727 would be group 2, so on and so forth...

Exact results should be:

31
21
215
7
6
40

with the given data

Answer 1

Try this out:

SELECT
  min(date(event_timestamp)) aDay,
  max(UNIX_TIMESTAMP(event_timestamp)) - min(UNIX_TIMESTAMP(event_timestamp)) secondDiff
FROM (
  SELECT event_timestamp, @eventGroup := @eventGroup + (event_count = 1) eventGroup
  FROM monitoring, (SELECT @eventGroup := 0) init
  WHERE hosting_id = 'c47204d5-d682-9ba4-b8b6-4fd26660651a'
  ORDER BY event_timestamp
) s
GROUP BY eventGroup

Output:

|             ADAY | SECONDDIFF |
|------------------|------------|
| October, 23 2013 |         31 |
| October, 23 2013 |         21 |
| October, 23 2013 |        215 |
| October, 23 2013 |          7 |
| October, 23 2013 |          6 |
| October, 24 2013 |         40 |

Fiddle here.