awk多个数组的使用

#!/bin/bash
awk 'BEGIN{printf "%-60s %-10s %-10s %-10s\n","url","count","time","avg_time"}
$NF ~ /^[-]/{
c1[$]+=$NF
c2[$]++}
END{
for(b in c2) {
    for(a in c1)
        if(a == b)
        printf "%-60s %-10s %-10s %-10s\n", a,c2[b],c1[a],c1[a]/c2[b]
        continue
}
}' log.txt
~           

注释：

c1统计每个url对应的总响应时间

c2统计每个url对应的总请求个数

每次循环，外围c2的值是不变的，内围c1处于变化中，当c1的url和c2相等时，则输出对应数据，同时跳出此次内循环，开启下一次新循环

获取每个url的平均响应时间

日志格式如下：

2014-06-20 16:47:43  [ resin-port-18057-51:333230 ] - [ INFO ]  [url] http://abc.com/list 254
2014-06-20 16:47:43  [ resin-port-18057-51:333230 ] - [ INFO ]  [url] http://abc.com/list 254

2014-06-20 16:47:43  [ resin-port-18057-51:333230 ] - [ INFO ]  [url] http://abc.com/list 254

2014-06-20 16:47:43  [ resin-port-18057-51:333230 ] - [ INFO ]  [url] http://abc.com/list 254

执行效果如下：

url                                                          　　　　　count      time       avg_time 
http://a.9tong.com/plat/active.do                           185         31            0.167568 
http://a.9tong.com/product/updatePic.do                 80          16465      205.812  
http://a.9tong.com/product/shopList.do                   1133       51663      45.5984