Linux下通过php对Mysql中的数据进行定时统计并入库

本文讲解Linux下对数据库的数据进行统计计算并将结果入库到另一个表中。本文分3部分进行讲解：第一部分介绍：linux下执行php的命令；第二部分介绍：通过脚本实现统计前几个月的数据并入库；第三部分介绍：如何实现每日凌晨对昨日的数据进行统计并入库。

1.Linux下执行php的命令

linux下执行php文件的命令为：php filename arguments。

即：php 文件名 参数

Eg：/usr/local/php/bin/php /home/zhaolincheung/insert_talkcountdata.php

2.通过脚本实现统计前几个月的数据并入库

例如，将之前60天内的数据统计入库。

首先，写一个shell文件talkcount_insert_db.sh，循环计算日期，并将该日期传给php文件，进行统计并入库。shell如下：

#!/bin/sh
for((i=1;i<=60;i++));
do
echo -e $i":";
day=`date -d "-"$i" day" +%Y-%m-%d`;
/usr/local/php/bin/php /home/zhaolincheung/insert_talkcountdata.php $day;
echo "";
done

而insert_talkcountdata.php是对某一天的数据统计并入库，如下：

<?php
　　include_once('/data/analys/DB/MyDB.class.php');
　　
//定义大类的编号
$cidArr = array(0,2,1,11,15,14,7,6,4,12,5,13,16,3,9,1001,1002);

$yesterday = $_SERVER["argv"][1]; //获取php执行文件的参数
echo $yesterday . "\n";

//遍历大类并入库
foreach($cidArr as $cid){
//统计大类
insert_db($cid,$yesterday);
}

//根据指定的大类和日期,计算该天的消息量并入库
function insert_db($category, $date){
$dateArr = explode("-", $date);
if(checkdate($dateArr[1], $dateArr[2], $dateArr[0]) == false){
return false;
}
//将日期转为时间戳
$thisdate = mktime(0, 0, 0, $dateArr[1], $dateArr[2], $dateArr[0]);
$nextdate = $thisdate + 3600 * 24;

$sql = "select talkid,count(*) as msgcount from table1 where postcategoryid=$category ";
$sql .= " and updatetime >= $thisdate and updatetime < $nextdate group by talkid order by msgcount desc";

$db = new MyDB();
$res = $db->query($sql);

//初始化关联数组
$mc["c20"]=$mc["c10"]=$mc["c6"]=$mc["c5"]=$mc["c4"]=$mc["c3"]=$mc["c2"]=$mc["c1"]=0;
//分段统计消息数对应的会话数
while(($result = $db->fetch_assoc($res)))
{
$msgcount = $result["msgcount"];
if($msgcount > 20)
$mc["c20"]++;
else if($msgcount > 10 && $msgcount <= 20)
$mc["c10"]++;
else if($msgcount >= 6 && $msgcount <= 10)
$mc["c6"]++;
else if($msgcount == 5)
$mc["c5"]++;
else if($msgcount == 4)
$mc["c4"]++;
else if($msgcount == 3)
$mc["c3"]++;
else if($msgcount == 2)
$mc["c2"]++;
else if($msgcount == 1)
$mc["c1"]++;
}

//insert table 'talkcount'
$sql_insert = "insert into talkcount(analysisdate,postcategoryid,c20,c10,c6,c5,c4,c3,c2,c1) values('";
$sql_insert .= $date . "',";
$sql_insert .= $category . ",";
$sql_insert .= $mc["c20"] . ",";
$sql_insert .= $mc["c10"] . ",";
$sql_insert .= $mc["c6"] . ",";
$sql_insert .= $mc["c5"] . ",";
$sql_insert .= $mc["c4"] . ",";
$sql_insert .= $mc["c3"] . ",";
$sql_insert .= $mc["c2"] . ",";
$sql_insert .= $mc["c1"] . ")";

echo $sql_insert . "\n";

$db->query($sql_insert);

$db->close();
}
?>

3.如何实现每日凌晨对昨日的数据进行统计并入库

另外，每日凌晨都需要将昨日的数据进行计算并入库。这里可以采用crontab -e命令，在计划任务中加入要执行的php文件。如：

0 5 * * * /usr/local/php/bin/php /data/analys/php/talkcount_Insertdb.php

表示每天凌晨五点执行/data/analys/php目录下的talkcount_Insertdb.php这个文件。

talkcount_Insertdb.php的内容如下：

<?php
　　include_once('/data/analys/DB/MyDB.class.php');
　　
//定义大类的编号
$cidArr = array(0,2,1,11,15,14,7,6,4,12,5,13,16,3,9,1001,1002);

$yesterday = date("Y-m-d",strtotime("-1 day")); //计算昨天的日期
echo $yesterday . "\n";

//遍历大类并入库
foreach($cidArr as $cid){
//统计大类
insert_db($cid,$yesterday);
}

//根据指定的大类和日期,计算该天的消息量并入库
function insert_db($category, $date){
$dateArr = explode("-", $date);
if(checkdate($dateArr[1], $dateArr[2], $dateArr[0]) == false){
return false;
}
//将日期转为时间戳
$thisdate = mktime(0, 0, 0, $dateArr[1], $dateArr[2], $dateArr[0]);
$nextdate = $thisdate + 3600 * 24;

$sql = "select talkid,count(*) as msgcount from table1 where postcategoryid=$category ";
$sql .= " and updatetime >= $thisdate and updatetime < $nextdate group by talkid order by msgcount desc";

$db = new MyDB();
$res = $db->query($sql);

//初始化关联数组
$mc["c20"]=$mc["c10"]=$mc["c6"]=$mc["c5"]=$mc["c4"]=$mc["c3"]=$mc["c2"]=$mc["c1"]=0;
//分段统计消息数对应的会话数
while(($result = $db->fetch_assoc($res)))
{
$msgcount = $result["msgcount"];
if($msgcount > 20)
$mc["c20"]++;
else if($msgcount > 10 && $msgcount <= 20)
$mc["c10"]++;
else if($msgcount >= 6 && $msgcount <= 10)
$mc["c6"]++;
else if($msgcount == 5)
$mc["c5"]++;
else if($msgcount == 4)
$mc["c4"]++;
else if($msgcount == 3)
$mc["c3"]++;
else if($msgcount == 2)
$mc["c2"]++;
else if($msgcount == 1)
$mc["c1"]++;
}

//insert table 'talkcount'
$sql_insert = "insert into talkcount(analysisdate,postcategoryid,c20,c10,c6,c5,c4,c3,c2,c1) values('";
$sql_insert .= $date . "',";
$sql_insert .= $category . ",";
$sql_insert .= $mc["c20"] . ",";
$sql_insert .= $mc["c10"] . ",";
$sql_insert .= $mc["c6"] . ",";
$sql_insert .= $mc["c5"] . ",";
$sql_insert .= $mc["c4"] . ",";
$sql_insert .= $mc["c3"] . ",";
$sql_insert .= $mc["c2"] . ",";
$sql_insert .= $mc["c1"] . ")";

echo $sql_insert . "\n";

$db->query($sql_insert);

$db->close();
}
?>