Shell Programming and Scripting

Is it possible to process multiple variables in a single sed command?
I have the following ksh with three variables and I want to search for all variables which start with "var" inside input.txt.
I tired "$var$" but it just prints out everyting in input.txt and does not work.

$ more test.ksh
#!/bin/ksh

var1=AAA
var2=BBB
var3=CCC

sed -n "/$var$/p" input.txt

$ more input.txt
DDD
CCC
EEE
AAA
BBB
$


My desired output is the below:

CCC
AAA
BBB

Any help will be appreciated.

Just use egrep....

Code:

#!/bin/ksh

var1=AAA
var2=BBB
var3=CCC

egrep "^(${var1}|${var2}|${var3})$" input.txt

Cheers
ZB

Thanks zazzybob.
However, for cases when there are like over 100 variables the code will be too long and will look ugly so I want way of some how using a wildcard character to process multiple variables. Is this possible?

The output order may be slightly different but you could do something like this...

Code:

#!/bin/ksh

var[0]=AAA
var[1]=BBB
var[2]=CCC

for i in ${var[@]}; do
  sed -n "/$i$/p" input.txt
done

exit 0

Cheers
ZB

use -e option of sed...

sed -n -e "/$var1$/p" -e "/$var2$/p" -e "/$var3$/p" input.txt

in this case file needs to be scanned for each and every search pattern

---->
better way would be to accumalate search patterns together and use egrep in a single shot

Code:

>cat input.txt
DDD
CCC
EEE
AAA
BBB

Code:

>cat search_pattern.txt
var1=AAA
var2=BBB
var3=CCC

Code:

>cat try.ksh
# !/usr/bin/ksh

pattern=""
for val in `sed -e 's/.*=//' search_pattern.txt`
do
   pattern=$pattern$val"|"
done

egrep -e $pattern input.txt
exit 0