sed'ing or greping until blank line

[Bill McNAMARA_1]

I use the following to get input for a script.

/usr/bin/sed -n '/'$1'/,$p' $3 |sed -n '1,'$2'p' > /tmp/temp

where $1 is the search string and $2 is the number of lines.

if I have an input file $3 like follows:

[XYZ]
hello = bonjour

[ABC]
RunDirectory = /opt/ABC/bin
RunString =/opt/ABC/bin/startABC
ProcessType = RS
RestartDelay = 15
Resources = DISK+MEM+LAN
StdoutFile = /tmp/ABC.out
StderrFile = /tmp/ABC.err
AutoStart = Yes
PreferredProcessor = 0

[BOB]
bob = robert

and search for ABC without specifying how many lines after I want the output file for the sed to include all the content up to the blank line.

I could use grep -n ABC in this case, but that won't work if there is an ABC1 and ABC2 so I'm hoping the sed line can be modified easily.

Thanks!
Bill

[Bill McNAMARA_1]

to add if I do the following

/usr/bin/sed -n '/'$1'/,$p' $3 > /tmp/temp

I get
# /sed.scr "ABC" /tmp/abc.conf "#"
[XYZ]
hello = bonjour

#[ABC]
#RunDirectory = /opt/ABC/bin
#RunString = /opt/ABC/bin/startABC
#ProcessType = RS
#RestartDelay = 15
#Resources = DISK+MEM+LAN
#StdoutFile = /tmp/ABC.out
#StderrFile = /tmp/ABC.err
#AutoStart = Yes
#PreferredProcessor = 0

#[BOB]
#bob = robert

and I don't want the [BOB] bit commented, just the [ABC] bit. ie. just up to the blank line

[James R. Ferguson · Accepted solution]

Hi Bill:

How about this:

# sed -n '/ABC/,/^$/p' myfile

...This will output everything between the two regular expressions, the second of which is a blank line.

Regards!

...JRF...

[Robin Wakefield]

Hi Bill,

Try:

/usr/bin/sed -n '/'ABC'/,/^$/p' $3 > /tmp/temp

Rgds, Robin

[Bill McNAMARA_1]

okay

here's my script again..
SEA=$1
# LEN=$2
FIL=$2
INS=$3
cp $FIL ${FIL}.orig

/usr/bin/sed -n '/'$SEA'/,/^$/p' $FIL > /tmp/temp

cat /tmp/temp | while read LINE
do
CHANGETO="${INS}${LINE}"
echo $LINE | sed 's/\[/\\\\[/g' | read LINE
sed "s:$LINE:$CHANGETO:g" ${FIL} > ${FIL}.bak
mv ${FIL}.bak $FIL
done

diff ${FIL} ${FIL}.orig
echo "file modified: backup at ${FIL}.orig"
rm /tmp/temp

Now I get a problem if the input file looks like this:

[XYZ]
hello = bonjour
PreferredProcessor = 0

[LMN]
hello = bonjour
PreferredProcessor = 1

[ABC]
RunDirectory = /opt/ABC/bin
RunString = /opt/ABC/bin/startABC
ProcessType = RS
RestartDelay = 15
Resources = DISK+MEM+LAN
StdoutFile = /tmp/ABC.out
StderrFile = /tmp/ABC.err
AutoStart = Yes
PreferredProcessor = 0

[BOB]
bob = robert

and I run the script to attempt to comment out the ABC section:

[XYZ]
hello = bonjour
#PreferredProcessor = 0

[LMN]
hello = bonjour
PreferredProcessor = 1

#[ABC]
#RunDirectory = /opt/ABC/bin
#RunString = /opt/ABC/bin/startABC
#ProcessType = RS
#RestartDelay = 15
#Resources = DISK+MEM+LAN
#StdoutFile = /tmp/ABC.out
#StderrFile = /tmp/ABC.err
#AutoStart = Yes
PreferredProcessor = 0

[BOB]
bob = robert


Any way around this!

Thanks,
Bill

[James R. Ferguson]

Hi (again) Bill:

OK. I prefer 'awk', if I understand your objective. Using your file:

#!/usr/bin/sh
#
SEA=ABC #...or SEA=$1
INP=/tmp/stuff #...or INP=$2
OUT=/tmp/stuff.out #...or OUT=$3
#
awk -v SEA=$SEA '$0~SEA {TOG=1}
/^$/ {TOG=0}
{if (TOG==1) {print "#" $0} else {print $0}}' $INP > $OUT
#.end.

Regards!

...JRF...

[James R. Ferguson]

Hi (again) Bill:

OK, here's a shorter piece of code that I believe accomplishes your objective. It's an alternative to my 'awk' script so you can take your choice. By example, using your data:

# sed -n -e '/ABC/,/^$/ s/^/\#/g' -e 'p' /tmp/stuff > /tmp/stuff.out

Regards!

...JRF...

Regards!

...JRF...

[Sridhar Bhaskarla]

Bill,

Doesn't this work for you?

sed '/abc/,/^$/s/^/#/' your_stuff

-Sri

[Sridhar Bhaskarla]

Ooops.. Sorry.. plase replace abc with ABC in the above.

-Sri

[Sridhar Bhaskarla]

One more ooops... Didn't verify my previous script. It is going to even comment out the
blank (^$) line that doesn't look good We need to exclude it from being commented out.


sed '/ABC/,/^$/ {/^$/!s/^/#/g;}' your_file

is going to comment out all the lines starting
with the line that has ABC till the blank line excluding the blank line.


-Sri

[Robin Wakefield]

Hi Bill,

How about this variation on a theme...

SEA=$1
FIL=$2
INS=$3
cp $FIL $FIL.orig

sed -e '/'$SEA'/,/^$/ s/^[^$]/'$INS'&/' $FIL > $FIL.bak
mv $FIL.bak $FIL

diff $FIL $FIL.orig
echo "file modified: backup at $FIL.orig"


Rgds, Robin.