in sed ,to get longest word

UNIX for Advanced & Expert Users
1

i want the longest word from the file using sed.
can any one help me in this case?

2

Sure it can be done a lot simpler....

awk ' BEGIN {
                X=0
            }
            {
                n = 1
                while ( n <= NF ) {
                    Y = length($n)
                    if ( Y > X ) {
                        X = Y
                        WORD = $n
                    }
                    n++
                }
             }
             END {
                     print WORD
                 } ' file
3 
awk '{ if ( length >a ) { a = length; b=$0 }}END{ print b }' filename
4

matrixmadhan - this will print the lonest line rather than the longest word though won't it?

5

Hi.

What is your definition of a word?

Why is the solution required to be with sed?

What have you tried so far?

cheers, drl

6

Or something like this with zsh/bash:

read< <(sort -rn <(for w in $(<file);do w="${w//[,.:\"\'()<>]}"
printf "%d %s\n" "${#w}" "$w";done));printf "%s\n" "${REPLY#* }"

For example:

bash 3.2.25(1)$ cat file
The Z shell (zsh) is a Unix shell that can be used as an interactive login shell 
and as a powerful command interpreter for shell scripting. Zsh can be thought of 
as an extended bourne shell with a large number of improvements, including some
of the most useful features of bash, ksh, and tcsh.
bash 3.2.25(1)$ read< <(sort -rn <(for w in $(<file);do w="${w//[,.:\"\'()<>]}"
printf "%d %s\n" "${#w}" "$w";done));printf "%s\n" "${REPLY#* }"
improvements
7 
awk 'BEGIN{ l=0}
   {  
	  for ( i=1 ;i<=NF;i++){
	     gsub(/[[:punct:]]/,"",$i)
   		 if (length($i) >l ) { 
			l=length($i)
			f=$i		 
		 }		 
	  }	  
   }
   END{
	  print l, f
   }' "file"

output:

# ./test.sh
12 improvements