how to merge similar lines in linux

问题:

I have a file test.txt on my linux system which has data in following format :

first second third fourth 10  
first second third fourth 20  
fifth sixth seventh eighth 10  
mmm nnn ooo ppp 10  
mmm nnn ooo ppp 20   

I need to modify the format as below -

first second third fourth 10 20  
fifth sixth seventh eighth 10 0  
mmm nnn ooo ppp 10 20  

I have tried following code

cat test.txt | sed 'N;s/\n/ /' | awk -F" " '{if ($1~$5){print $1" "$2" "$3" "$4" "$8} else { print $0 }}'

But this is not giving required output. When there is a line which doesn't have a similar line below it,this command fails. Can u suggest me any solution for this??

回答1:

Here is one way to do it:

awk ' {
  last=$NF; $NF=""
  if($0==previous) {
    tail=tail " " last
  }
  else {
    if(previous!="") {
      if(split(tail,foo)==1) tail=tail " 0"
      print previous tail
    }
    previous=$0
    tail=last
  }
}
END {
    if(previous!="") print previous tail
}
'

回答2:

Perl solution:

perl -ne '/^(.*) (\S+)/ and push @{ $h{$1} },$2 }{ print "$_ @{$h{$_}}\n" for keys %h' < test.txt

回答3:

Reuse of my solution (J4F)

cat file.txt | sort | while read L;
do
  y=`echo $L | rev | cut -f2- -d' ' | rev`;
  {
    test "$x" = "$y" && echo -n " `echo $L | awk '{print $NF}'`";
  } || 
  {
    x="$y";echo -en "\n$L"; 
  };
done