两个shell例子

1.在ping前面加时间
    while :;do ping -c 1 172.17.39.251|awk '/ttl=/' | sed "s/^/`date +%Y-%m-%d\|%T` /";sleep 1;done

2.消除相同的文件名
./2006/20060721b/p044559_47.txt
./2006/20060728a/p044559_47.txt
./2007/20070606News/p044559_47.txt
./Olympic City/Shanghai/Cityscape/p044559_47.txt
./2006/20060721b/p053826_01.txt
./Olympic City/Shanghai/Cityscape/p053826_01.txt
./2006/20060719News/Gome/pau52839_02.txt
./2006/20060725a/pau52839_02.txt
./2006/20060726News/TyphoonKeami/pau90550_12.txt
./2006/20060728News/SMS/pau90550_12.txt
./2006/20060714News/Motorola/dfzb20050113_056.txt
./2006/20060727News/Huawei/dfzb20050113_056.txt
./2006/20060714News/Motorola/p055816_15.txt
./2006/20060727News/Huawei/p055816_15.txt
./2006/20060714News/Motorola/p060766_04.txt
./2006/20060727News/Huawei/p060766_04.txt
./2006/20060814a/Motorola/p060766_04.txt

文本中的内容是文件所在的路径
怎么实现如果相同文件名则删除,只留其中的一份
比如
./2006/20060721b/p044559_47.txt
./2006/20060728a/p044559_47.txt
./2007/20070606News/p044559_47.txt

则删除
./2006/20060728a/p044559_47.txt
./2007/20070606News/p044559_47.txt

只留
./2006/20060721b/p044559_47.txt


    awk -F\/ '!a[$NF]++'  file 相当于
awk -F\/ '!a[$NF]++{print $0}'  file

!a[..]++是一个pattern， 只有匹配该pattern时，才执行相应的action，即print $0.
第一次时, a[...]为0 所以!a[..]为真，执行print $0. 再碰到相应的文件名a[..]不为0，!a[..]为假，所以....