UNIX for Dummies Questions & Answers

Little bit confusing while using awk
In Sed while pattern search we can use "(double quotes)
i mean

Code:

$a=hello
$cat file.txt |sed -n "/$a/p"

this thing work fine But if i use it in awk it's not working How could i do the substitution of pattern by a variables and the result as the desired position mean the 2nd and 4th location data only not the whole line. if i go for this

Code:

$cat file.txt | awk "/$a/{print $2,$4}"

in $2,$4 at ,(comma) the syntax error is coming.

Thank & Regard's
Posix
-------------------
Luv to learn awk and sed

The dollar sign '$' tells the shell that the characters trailing it form variable name. Besides the usual variables, there are also special ones like $@, $*, and $0 to $9.

Your sed command gets transformed like this:

Code:

cat file.txt |sed -n "/$a/p"
cat file.txt |sed -n /hello/p

With the awk command you've given, the shell tries to interpolate the $2 and $4 with the appropriate positional parameters, which probably aren't set. So the transformation would look like this:

Code:

cat file.txt | awk "/$a/{print $2,$4}"
cat file.txt | awk /hello/{print ,}

But fear not, there is help. awk supports the -v switch, which allows one to set awk variables from outside the awk program. So if you rewrite the line as

Code:

awk -vsearch=$a '$0 ~ search {print $2,$4}' file.txt

everything should work. And you're even saving on time since you don't have to run cat unnecessarily.