sed command

[Pavitra]

I Have attached the file. I need the file in the format
value

[Cédric MELLON]

Hi,

you can do that with sed and awk
sed : add a separator character before the string "value" => "!value" for exemple
awk : use this character as begin of line and the carriage return as separator of fields

Hope that will helps,
Cedric.

[Peter Godron]

Hi,
not in sed/awk, but old-fashioned shell ;-)

#!/usr/bin/sh
a=1
while read record
do
if [ $a -eq 1 ]
then
value="$record"
a=`expr $a + 1`
else
count=`echo "$record" | awk '{print $3}'`
count2="`echo $count2` $count"
if [ $a -le 3 ]
then
a=`expr $a + 1`
else
echo $value $count2
a=1
count2=""
fi
fi
done < a.lis


Please also read:
http://forums1.itrc.hp.com/service/forums/helptips.do?#33 on how to reward any useful answers given to your questions.

[Cédric MELLON · Accepted solution]

Here is your command, tested ;-p

cat test | sed 's/value/!value/g' | awk 'BEGIN { RS = "!" ; FS = "\n" } {print $1 " " $2 " " $3 " " $4}'

Cedric.

[Dennis Handly]

You can use awk. (Your example input contains an initial blank line that causes problems. Did you really want it? The scripts below assume it isn't there.)

awk '
{
s1 = $0
getline
s2 = $0
getline
s3 = $0
getline
print s1, s2, s3, $0
}' 284910.txt

The following sed command will also work:
sed -n 'N;N;N;s/\n/ /g;p;d' 284910.txt

[Peter Nikitka]

Hi,

I assume you only want to print the real values, no the whole text.
I added a header - if you don't need it, drop the 'BEGIN' section.

awk 'BEGIN {printf("%s\t%s\t%s\t%s\n","value","insert","update","delete")}
NF==1 {if(val) printf ("%s\t%d\t%d\t%d\n",val,ins,upd,del);val=$1}
/^Rows inserted/ {ins=$3}
/^Rows updated/ {upd=$3}
/^Rows deleted/ {del=$3}
END {if(val) printf ("%s\t%d\t%d\t%d\n",val,ins,upd,del)}' yourfile

mfG Peter

[Pavitra]

Hey,
Thanz Peter, Cedric,Dennis and Peter..
I got the required output..:-)

[Peter Nikitka]

Hi,

nice to hear, that our solutions were valuable! Because you are new to this forum - Welcome! - I want to put your attention to its points system:
http://forums1.itrc.hp.com/service/forums/helptips.do?#28

mfG Peter

[Sandman!]

Yet another way to do the same thing using ex(1):

ex -s file < g/^value/ j 3 | p
EOF

[Dennis Handly]

>Sandman!: g/^value/ j 3 | p

Thanks, I was never able to figure out how to join multiple lines in vi until I saw this and reread the man page.

Though in visual mode, having to type ":j3" is not much easier than "JJJ"

[Pavitra]

thanx

[Peter Nikitka]

Hi Pavitra,

to make it easier for others to find valuable solutions for their problems, please rate the answers to your questions.

Here you'll find a description about this:
http://forums1.itrc.hp.com/service/forums/helptips.do?#33

mfG Peter