Home / ubuntu / Questions / 691960
In Process
Roboman1723
Asked: 2015-10-31 07:47:02 +0800 CST

Reverse string matching

  • 772

So basically what I want to do is use a grep,sed or awk command (I'm not picky) that can do the following. I know how to go from Vuln 1 to OS1 but not vice versa.

Input:

test.txt 
Vuln 1
 OS1
 OS2
 OS3

Vuln 2
 OS4
 OS5

Output (heavy pseudocode):

$ grep "OS1" test.txt
Vuln 1
bash

3 Answers

  1. I guess, this is what you need:

    perl -ne 'chomp; $h=$_ if $_=~/^V/; print $h,"\n" if $_=~/OS1/' test.txt

    Explanation:

    1. perl -ne reads the input linewise
    2. chomp removes newline at the end (not necessary, but I think, it is cleaner)
    3. The variable $h contains the content of the line, if it starts with V
    4. $h is printed if the current line contains OS1
    • 1

  2. In this solution we use GNU AWK.

    awk -F"\n" -v RS="" '/OS1/ {print $1}' test.txt

    Explanation

    • we take empty line as a record separator (RS="")
    • fields are separated by newlines (-F"\n")
    • the program is: if we find "OS1" in the record, print the first field (/OS1/ {print $1})
    • 1

  3. You can do:

    grep -Pzo '\n?\KVuln 1(?=\n\s*OS1\n)' test.txt

    • We are using PCRE (-P) and ASCII NUL as input line delimeter (-z) instead of \n so that we can match \n literally

    • \n?\K will match an optional newline before our desired Vuln 1, \K will discard the match

    • (?=\n\s*OS1\n) is the zero width positive lookahead pattern ensuring that Vuln 1 is followed by OS1 in a \n separated line.

    Example:

    $ cat test.txt 
Vuln 1
 OS1
 OS2
 OS3

Vuln 2
 OS4
 OS5

$ grep -Pzo '\n?\KVuln 1(?=\n\s*OS1\n)' test.txt 
Vuln 1
    • 0