Hi Experts

I am facing a problem while using sed to replace a multiline string

I have sed command stored in a file say command

command file has the following content

s/"Owner"/"Owner_T"/g
s/"Contract
Number"/"Contractnumtest"/g

I want to replace
"Contract
Number" string with "Contractnumtest"
the file with name say "file" which i have to edit is something like this

"Owner"
"Contract"
"Contract
Number"

when i try to use sed to edit by doing this

sed -f commands $file > file_outptut1.txt

I am getting errors like
sed: command garbled: s/"Contract

I am using Solaris as OS.

Is it possible to replace the multiline string using sed or sed commands have to be end in one line???

any solution/suggestion/pointers to solve this problem are welcome

thanks in advance for helping

You need to consult the sed&awk bible! Here:
http://www.unix.org.ua/orelly/unix/sedawk/ch06_01.htm

Basically you should be able to use what's called the "hold space" to recognise and remember stuff from multiple lines.

HTH

Hi,

Your command file should look something like this:

s/"Owner"/"Owner_T"/
/"Contract$/{
N
s/"Contract\nNumber"/"Contractnumtest"/
}

Nick_Battle gave good advice: All the info needed is in the Sed&Awk book.

Hi.

I give you no extra magic, however, a rule that works for me is to remember that sed deals with single-lines unless you do some extra work.

I could not get the URL that Nick posted to work ( page deleted ). If you can read Russian, perhaps some of the links at that top page might help.

There are some other places to look. The web page http://sed.sourceforge.net/ has a number of links that may help.

It is difficult to imagine a better reference than O'Reilly's sed and awk. I use the 2nd edition (page 107 ff for your circumstance); you might be able to pick up a used version someplace of either edition 1 or 2. I think sed is very useful in many situations, however, perl has probably taken over much of the new work for which sed had been used.

Best wishes ... cheers, makyo