Search Results

replace

> output.txt

to

>> output.txt

Try...awk '{a[$1]+=$2;b[$1]+=$3}END{for(c in a)print c,a[c],b[c]}' file1

Duplicate thread.
https://www.unix.com/shell-programming-scripting/154633-awk-programming.html

I don't know what gave you the impression that sed would do a faster job than grep or awk for the same matter.

The syntax for what you want is:
sed -n '/matching_pattern/ p' filename

Which it...

one way:


# sed -e 's/,/ /g' -e 's/\"\([0-9]*\) \([0-9]*\)\"/\1\,\2/g' infile
0 1 0 87 0 0 6,87 170,03 0 43,5 0 0 0 0 6,87 126,53


HTH

Another way, which is pretty similar to Tytalus' approach:
$ echo '0,1,0,87,0,0,"6,87","170,03",0,"43,5",0,0,0,0,"6,87","126,53"' |\
> sed 's/"\([0-9]*\),\([0-9]*\)"/\1.\2/g;s/,/ /g;s/\./,/g'
0 1...

The easiest way to solve that problem is to install gawk ;) I think it is included in Solaris 10 Companion CD.

You could also use Perl:
perl -pe 's/,/ /g;s/\"([0-9]+) ([0-9]+)\"/\1,\2/g' infile

nawk 'FNR==NR {f1[$1];next} $1 in f1' file1 file2