Very Computer

Hi
I have a file.txt like this

aword1 aword2 aword3 XXX = aword5 YYYY aword1 aword3 aword3 aword5
aword7 aword2 aword8 aword9 aword10 aword11 aword12 aword13 aword16
aword17

I have to extract 'aword5' using simplest combination of sed/awk. I
believe it can be done with sed only but wasn't able to find exact
"magic" expression.

One expression works though:

sed -n "s/XXX.=./#/p" ./file.txt |awk -F"#" '{ print $2 }' |awk ' {
print $1 }'

I am looking for simpler/faster version... and I don't like the
possibility that # may be already present. Help appreciated!

Thanks
Andrew

    From the code you posted, I believe what you really want is
    "the word following 'XXX = '"; is that right?

    With a POSIX shell, using only cat:

var=`cat file.txt`
var=${var#*XXX = }
echo ${var%% *}

    With bash2, no cat:

var=`< file.txt`
var=${var#*XXX = }
echo ${var%% *}

    With sed:

sed -n "s/.*XXX = \([^ ]*\) .*/\1/p" file.txt

--
    Chris F.A. Johnson                        http://cfaj.freeshell.org
    ===================================================================
    My code (if any) in this post is copyright 2003, Chris F.A. Johnson
    and may be copied under the terms of the GNU General Public License

Given the contents you say are in your file, you should be able to do
it without either sed or awk. Try something like this

#!/bin/ksh
f='aword1 aword2 aword3 XXX = aword5 YYYY aword1 aword3 aword3 aword5
aword7 aword2 aword8 aword9 aword10 aword11 aword12 aword13 aword16
aword17'

set -- $f
while [ "$1" != XXX ];do
  shift
done

if [ $# -gt 2 ];then
  shift 2
  echo $1
else
  echo "Not found"
fi

Joe

Quote:>Hi
>I have a file.txt like this

>aword1 aword2 aword3 XXX = aword5 YYYY aword1 aword3 aword3 aword5
>aword7 aword2 aword8 aword9 aword10 aword11 aword12 aword13 aword16
>aword17

>I have to extract 'aword5' using simplest combination of sed/awk. I
>believe it can be done with sed only but wasn't able to find exact
>"magic" expression.

>One expression works though:

>sed -n "s/XXX.=./#/p" ./file.txt |awk -F"#" '{ print $2 }' |awk ' {
>print $1 }'

>I am looking for simpler/faster version... and I don't like the
>possibility that # may be already present. Help appreciated!

Hi Andrew,

The simplest is probably to use perl;

perl -ne '/(^|\s)XXX\s+=\s+(\S+)/ && print $2;'

I modified your RE to match only when 'XXX' and '=' are seperate
words :)

awk '{for(i=0;++i<NF-1;)if($i=="XXX"&&$(i+1)=="=")print $(i+2)}

This 'sed' script actually uses a marker character '|' like
your hash however at that stage the '|' in the original text
is mapped to a metacharacter ':2',

sed '
  s/:/:1/g
  s/|/:2/g
  s/^XXX[       ]*=[    ]*\([^  ]*\)/|\1|/
  s/[   ]XXX[   ]*=[    ]*\([^  ]*\)/|\1|/
  /|/!d
  s/[^|]*|\([^|]*\).*/\1/
  s/:2/|/g
  s/:1/:/g
  '

byebye
l

Quote:

>Thanks
>Andrew

--

using 'sed' alone:
sed -e '/=/s/^.* XXX = \([^ ]*\) YYYY .*/\1/' ./file.txt

using 'sed/cut' combo:
sed -e '/=/s/ YYYY / = /' | cut -d= -f2