File searching and proximity search

Question

I've been looking at grep a file, but show several surrounding lines?

I'm using a bash terminal, looking for a file which

• Has both path and redirect on any one line

• Has flash on a nearby line, within five lines from the first

In this possible with grep, ag, Perl, sed or any tool you guys know of?

Answer 1

The easier filter is the one with "flash". It is also good to do it first, so that the more expensive pattern matching is done in the subset of matched files.

For this, just say:

grep -RH -C 5 "text" *

This will recursively (-R) look for the pattern "text" and print the name of the file (-H) when this happens. Also, it will print the surrounding 5 lines (-C 5). Just change 5 with a variable if you want so.

Then it is time to use awk to check two patterns with:

awk '/pattern1/ && /pattern2/ {print FILENAME}' file

This is useful as awk is quite good on multiple patterns matching.

Since we do not have the filename but a stream on the form filename:flash, etc, we can have a basic Bash loop to handle the result from grep:

while IFS=":" read -r filename data;
do
    awk -v f="$filename" '/path/ && /redirect/ {print f}' <<< "$data"
done < <(grep -RH -C5 "text" *)

Answer 2

ack -A6 -B4 'path.*redirect|redirect.*path' FILES | grep flash

outputs lines that contain the pattern flash in the 4 lines before or 6 lines after the lines in the files FILES that contain the patterns path and redirect along with the filename and the line number of the line containing flash.

Without the ack command (or the egrep command, which will also work), you can rephrase this as two grep commands

(grep -A6 -B4 'path.*redirect' FILES ; grep -A6 -B4 'redirect.*path' FILES) |
    grep flash

Answer 3

This is a bit more complicated than it seems, because you're looking for words in rough proximity.

So I'd probably tackle it a bit like this:

#!/usr/bin/env perl

use strict;
use warnings;

my $buffer_limit = 5; # +/- 5

my @buffer; 

my $first_flag;
my $second_flag; 

#iterate stdin or files specified on command line
while ( my $line = <> ) {

   #test first condition
   if ( $line =~ m/path/ and $line =~ m/redirect/ ) { $first_flag++; };
   #test second condition
   if ( $line =~ m/flash/ ) { $second_flag++; };

   #if either is true - match has been seen recently. 
   #save the line in the buffer. 
   if ( $first_flag or $second_flag ) { 
         push @buffer, $line
   }
   #if both are true, we print (and flush the buffer)
   if ( $first_flag and $second_flag ) { 
       print "Match found up to line $.:\n";
       print @buffer;
       @buffer = ();
       $first_flag = 0;
       $second_flag = 0; 
   }
   #exceeding limit means that both matches haven't been seen in proximity. 
   if ( @buffer > $buffer_limit ) { 
      @buffer = ();
      $first_flag = 0;
      $second_flag = 0;
   }
}

We use a rolling 5 line buffer. We start capturing when we hit one or other 'match' and we print/flush if we hit the second match. And then empty the buffer if we exceed 5 lines.