mjhjmtu

I have a file like the one below. I want to return the value to the right of info2. Secondly I would want to return the value of info5 in the same manner.

FILE:

info1 218, info2 111, info3 415, info5 done,

With awk:

$ awk 'BEGIN RS="," $1 ~ /info[25]$/ print $2 ' input
111
done

We set the input record separator (RS) to the comma, and then for each record in which the first field contains info followed by either a two or a five (and nothing else), print the second field.

A slight improvement on choroba’s answer:

sed -n 's/.*info2 ([^,]*).*/1/p' file

• 
  -n means “don’t print anything
  except when specifically commanded to do so”. 
  This is (somewhat) equivalent to choroba’s use of !d
  to delete (ignore) lines that don’t match info2.


• The guts of the command is very similar to choroba’s,
  except it captures a sequence of any characters except commas
  (not just numerals), so it can extract the done from info5 done,.


• The command ends with p (print),
  because otherwise, with the -n, nothing would be printed.

This has the advantage that
it doesn’t require specifying the infoN label twice. 
(And it doesn’t generate undesired results
if a line contains info20, info21, etc.)

A two step solution:

tr ',' 'n' <file | sed -n 's/^ *info[25] *//p'

The tr transforms the example text into

info1 218 
 info2 111 
 info3 415 
 info5 done 

and the sed script removes the info2 and info5 strings (and any surrounding whitespace), leaving the wanted pieces of text. The other lines are discarded.

Result:

111
done

This is better done with awk, as in DopeGhoti's solution.

An awk variation based on my tr approach:

tr ',' 'n' <file | awk '/info[25]/ print $2 '

You can use sed:

sed '/info2 [0-9]/!d;s/.*info2 ([0-9]*).*/1/' file

• 
  /info2/ is an "address", any line containing info2 is considered to match the address. !d means "delete lines outside of the address", so only lines containing info2 will be processed by the next command.


• The substitution uses a capture group to remember the number after info2, and the whole line (everything before info2, info2 itself, the number, and anything following the number) is replaced by the number.

If you need to capture "done" after info5, you can't use [0-9]*, but you can replace it by [^,]*, i.e. non-comma.