Shell Programming and Scripting

hi all

is possible to pass shell (bash) variable to sed like it is in awk?

example:

awk script is storred in awk.awk file and I am passing variable called var to this file.

Code:

$ cat awk.awk
{if ($5==var) print $0}

so it works when i issue

Code:

$ bash_var=24
$ ls -l | awk -v var="$bash_var" -f awk.awk
-rw-r--r-- 1 cepi cepi     24 2010-03-20 13:01 awk.awk

Is possible to do similar thing to sed ?

Code:

$ cat sed.sed
sed "$res""s/$/(-,"$name",time2change.co.uk)/"

$res should be number of line which sed will be applied to
$name should be some replace string

so i will run it like

Code:

cat file | sed -f sed.sed     # variable definition will go somewhere here

Thank you

---------- Post updated at 09:42 AM ---------- Previous update was at 09:41 AM ----------

If is not this possible what is your favorite way to pass bash variables to sed ?
I wanted to try this way because sed program could be stored in separate file (so it would not be need to use " and so avoiding escaping and white space interpretation)
Thank you

Here is generic method to use dynamic templates.

Code:

#!/bin/ksh or bash or ...
parse_file()
{
 eval echo  "\"$(cat $1 | sed 's+\"+\\"+g'   )\""
}

###########
replace=newvalue
export replace
parse_file sed1.sed  > $$.tmp
sed -f $$.tmp sed1.txt
rm -f $$.tmp 2>/dev/null

sed1.sed file

Code:

s/$replace/replaced/g

If you have only a sed rule, use something like this

Code:

rule="${res}s/$/(-,$name,time2change.co.uk)/"
sed -e "$rule" somefile