mysql中的多次分组。（表中重复的信息，处理方式）

<![CDATA[
	  SELECT
  a2.plateNo,v.plateColor,d.name as depName,COUNT(a2.plateNo)*60*24 AS onlineTime,(TIMESTAMPDIFF(DAY,DATE_FORMAT(MIN(a2.createDate), '%Y-%m-%d'),DATE_FORMAT(MAX(a2.createDate), '%Y-%m-%d'))+1)*60*24 as totalTime
  ,COUNT(a2.plateNo)/(TIMESTAMPDIFF(DAY,DATE_FORMAT(MIN(a2.createDate), '%Y-%m-%d'),DATE_FORMAT(MAX(a2.createDate), '%Y-%m-%d'))+1)*100 as onlineRate
FROM
  (SELECT
    o.plateNo,o.createDate
    FROM
    onlinerecord o
  WHERE alarmType = "GpsOnline"
  GROUP BY plateNo,
    DATE_FORMAT(createDate, '%Y-%m-%d')) a2,department d,vehicle v
]]>
		<where>
			LEFT JOIN (SELECT * FROM vehicle GROUP BY plateNo) v ON a2.plateNo = v.plateNo

			LEFT JOIN department d ON v.depId = d.depId
mysql语句解释
（1）from中有一个表示a2
SELECT
 o.plateNo,o.createDate
  FROM
  onlinerecord o
  WHERE alarmType = "GpsOnline"
  GROUP BY plateNo,
  DATE_FORMAT(createDate, '%Y-%m-%d')

进行2次分组。 第一次分组按车牌号进行分组；
第二次分组按日期（精确到日）进行分组；
下面的left join(目的是：以a2为主体)；
其中对vehicle进行按plateNo(其中的vehicle表中大量重复的信息。这样分组后，即可解决)

<![CDATA[]]>自己百度。这里我觉得可以去掉。