Sort complex data

Shell Programming and Scripting
1

Hi,

Can someone here help sorting the following data in numeric order?

INPUT:
FIRST abc(3) def(13) fgh(1) ijk(6) abc(2)
SECOND dfe(10) abc(4) hij(19) tlm(1) hij(1) hub(10) abc(1) fed(3)
OTHERS hij(10) mok(4) bub(19) hij(1) abc(2) abc(15) abc(1) hij(3)

OUTPUT:
FIRST def(13) ijk(6) abc(3) abc(2) fgh(1)
SECOND hij(19) dfe(10) hub(10) abc(4) fed(3) abc(1) hij(1) tlm(1)
OTHERS bub(19) abc(15) hij(10) mok(4) hij(3) abc(2) abc(1) hij(1)

Thanks in advance for your help!!!

2

That output is tricky, since it seems to be sorting in two different directions, alphabetically but reverse-numerically... commandline sort can't do that, you'd be writing your own shell-based sorting routines. Is the form of the output really that strict? What's the data for?

3

Are you sorting each line by the number in () in descending order?

4

The data is from CVS log and I only care about the number, not alphabetic.
All I want to do is sort numers in each row.

Thanks!

5

Yes. Need to sort numbers in descending order in each line.

Thank you!

6

Here's some code that could probably be more efficient but I think does what you want:

#!/bin/sh

while read TITLE LINE
do
        ARR=( $LINE )
        for ((N=0; N<${#ARR}; N++))
        do
                if [[ -z "${ARR[$N]}" ]]
                then
                        break
                fi

                OLDIFS="${IFS}"
                IFS="()"
                VAL=( ${ARR[$N]} )

                printf "%s %s\n" ${VAL[1]} ${VAL[0]}
                IFS="${OLDIFS}"
        done | sort -rn | (
                echo -n $TITLE
                while read I STR
                do
                        echo -n " ${STR}(${I})"
                done
                echo    )
done

exit 0

$ echo "FIRST abc(3) def(13) fgh(1) ijk(6) abc(2)
SECOND dfe(10) abc(4) hij(19) tlm(1) hij(1) hub(10) abc(1) fed(3)
OTHERS hij(10) mok(4) bub(19) hij(1) abc(2) abc(15) abc(1) hij(3)" | ./sorter.sh
FIRST def(13) ijk(6) abc(3) abc(2) fgh(1)
SECOND hij(19) hub(10) dfe(10) abc(4) tlm(1) hij(1) abc(1)
OTHERS bub(19) abc(15) hij(10) mok(4) abc(2) hij(1) abc(1)
$
7

It failed at line "ARR=( $LINE )".
syntax error at line 5: `ARR=' unexpected

8

need_help - what shell are you using? show the result from

echo $SHELL
9

Try /bin/bash instead of /bin/sh. I keep forgetting that plain sh doesn't always support "advanced" things like.. sigh.. arrays..

10 
> cat file167
FIRST abc(3) def(13) fgh(1) ijk(6) abc(2)
SECOND dfe(10) abc(4) hij(19) tlm(1) hij(1) hub(10) abc(1) fed(3)
OTHERS hij(10) mok(4) bub(19) hij(1) abc(2) abc(15) abc(1) hij(3)

> cat sort167.sh
#!/usr/bin/bash

while read header data
   do
#   echo $header
#   echo $data
    revdata=`echo $data | tr " " "\n" | sort -rn +0.4 | tr "\n" " "`

    echo $header $revdata

done <file167

> sort167.sh
FIRST def(13) ijk(6) abc(3) abc(2) fgh(1)
SECOND hij(19) hub(10) dfe(10) abc(4) fed(3) tlm(1) hij(1) abc(1)
OTHERS bub(19) abc(15) hij(10) mok(4) hij(3) abc(2) hij(1) abc(1)
11

With Perl:

perl -lane'
    print join " ", shift @F, sort {
        ($aa) = $a =~ /(\d+)/;
        ($bb) = $b =~ /(\d+)/;
        $bb <=> $aa
      } @F
  ' infile
12

hi, i think perl should be a right and easy solution.

#!/usr/bin/perl
sub sub_sort{
	my @t1=split("[(|)]",$_[0]);
	my @t2=split("[(|)]",$_[1]);
	if($t1[1] == $t2[1]){
		return $t1[0] cmp $t2[0];
	}
	else{
		return $t2[1] <=> $t1[1];
	}
}
open FH,"<a.txt";
while(<FH>){
	chomp;
	my @temp=split(" ",$_);
	print $temp[0]," ", join " ",sort { sub_sort($a,$b) } @temp[1..$#temp];
	print "\n";
}
13

Nice catch summer_cherry, I did not notice that the OP wants to sort twice (first numeric, then ascii/utf8), so I need to modify the code:

perl -lane'
    print join " ", shift @F, sort {
        ($aa) = $a =~ /(\d+)/;
        ($bb) = $b =~ /(\d+)/;
        $bb <=> $aa || $a cmp $b
      } @F
  ' infile