Sed, awk & piping

Question

Hey all,
I'm wanting to look for a specific line and word which I have been able to with the following command
sed -n '3p' | awk {'print $2'}
I then couldn't really figure out how to properly pipe the awk result out so I made the whole command into a variable like so:
var="$(sed -n '3p' | awk {'print $2'})"
echo $var;
This give the same output.
Now one of two things. So is there a way to properly pipe the result into another sed command with out having to set it as a variable
OR
I tried using said variable to use the sed command again separately, only this time to replace the value i found.
with sed 's/$a/some_string/g' FILE  (the file is a (dot) file if it matters)
No matter how I seem to do this I can't seem to get the variable to work with the search and replace tool of SED.
Any help would be much appreciated. Let me know if sed or ed would be better to use.

Answer 1

So is there a way to properly pipe the result into another sed command with out having to set it as a variable
just put another vertical bar and the next command, such as the following:
sed -n '3p' input.file | awk '{print $2}' | next_command | and_the_next | etc...
var="$(sed -n '3p' | awk '{print $2}')"
echo $var;
If all you wanted to do was print the output from 'awk', then you shouldn't need to do anything special beyond:
sed -n '3p' input.file | awk '{print $2}'
as awk sends its output to standard out, just like 'echo'.  Of course if you want to do "Other" things to 'var', then it can be useful to capture the output, but to just display it via 'echo' all you need to do is let 'awk' send it to standard output as it normally does.
You 'awk' notation is wrong.  You have { 'print $2' }, when what you should have is ' { print $2 } '.  That is to say, the { print $2 } should be inside the single quotes.
If I understand your "sed -n '3p' input.file", you are printing the 3rd line of the input, then capturing the 2nd field via 'awk'.  You could do all of this with just awk
awk 'NR==3 { print $2; exit }' input.file
Where NR is the current record (line) number, so when NR == 3, it is the 3rd line.  The 'exit' stops the awk script after you get what you were after, so awk does not read the rest of the file.
If you want to capture the 2nd field for additional use in the script, such as a substitute value, then it is good to capture the field in a variable
var=$(awk 'NR==3 {print $2}' input.file)
sed -e "s/$var/some_string/g" input.file >output.file
Important things to note.  I used double-quotes "..." so that the shell will substitute the contents of $var into the 'sed' command.
HOWEVER, I use single-quotes '...' around the 'awk' commands because I DO NOT want the shell to see the $2 and the single-quotes protect the $2 from the shell, so that 'awk' can see the $2 and do the 'awk' things with $2.
The shell will perform variable substitutions inside double-quotes "...".  The shell will NOT do variable substitution inside single-quotes '...'
If you would prefer to use var=$(sed -n '3p' input.file | awk '{print $2}') that is perfectly OK, just less efficient, but if you understand it, then it works.