October 30, 2010

Perl Sort Hash by Value

By Mohammed Abualrob Algorithms and Data Structures 0 Comments

Problem

Given a log file similar to Linux syslog format as in the following example:

Jan 4 03:15:01 pla kernel:pla: Some error message

1	Jan 4 03:15:01 pla kernel:pla: Some error message

Write a Perl script to print the top 5 most repeated error messages.

Solution

Read the file line by line then split each line using “:” as separator then use the error message token as the key in a hash table and the error message count as the corresponding hash value. At the end you will get all error messages and their counts stored in the hash table. After that you need to sort the hash table by value not by key in decreasing order. The final step is to loop through the hash and print the first 5 values.

Code

Here is the code in Perl

#!/perl/bin/perl

# Command line arguments
my $Args = "@ARGV";

# log file name
if ($Args =~ /--log (S+)/)
{
   $log = $1;
}

# Hardcoded for testing purpose only
$log = "/perl/test/Log.txt";

# Check tool usage
if (!(defined($log)))
{
   print "\n";
   print "Usage: question_two.pl [--log ]\n";
   print "\n";
   exit;
}

print "Please wait...\n\n";

# Open file for reading
open (F, $log) or die "can not open file: $!";

# Hash table:
# key: log message
# value: count
my %MessageCount = ();

# While there are lines to read
while ()
{
   # Remove EOL
   chomp($_);

   # Split current line into tokens
   # using : as separator
   # Message is token 5
   @tokens = split(/:/, $_);

   # Hash message while incrementing
   # its counter
   $MessageCount{$tokens[4]}++;
}

# Close file
close(F);

# Counter to 5
my $i = 0;

# Sort hash by value (recall: not by key)
foreach $key (sort{$MessageCount {$b} <=> $MessageCount {$a}} keys %MessageCount)
{
   $i++;

   print "$key : $MessageCount{$key}\n";

   # Exit after printing the 5th most repeated message
   if ($i == 5)
   {
      last;
   }
}

#!/perl/bin/perl

# Command line arguments

my $Args = "@ARGV";

# log file name

if ($Args =~ /--log (S+)/)

{

$log = $1;

}

# Hardcoded for testing purpose only

$log = "/perl/test/Log.txt";

# Check tool usage

if (!(defined($log)))

{

print "\n";

print "Usage: question_two.pl [--log ]\n";

print "\n";

exit;

}

print "Please wait...\n\n";

# Open file for reading

open (F, $log) or die "can not open file: $!";

# Hash table:

# key: log message

# value: count

my %MessageCount = ();

# While there are lines to read

while ()

{

# Remove EOL

chomp($_);

# Split current line into tokens

# using : as separator

# Message is token 5

@tokens = split(/:/, $_);

# Hash message while incrementing

# its counter

$MessageCount{$tokens[4]}++;

}

# Close file

close(F);

# Counter to 5

my $i = 0;

# Sort hash by value (recall: not by key)

foreach $key (sort{$MessageCount {$b} <=> $MessageCount {$a}} keys %MessageCount)

{

$i++;

print "$key : $MessageCount{$key}\n";

# Exit after printing the 5th most repeated message

if ($i == 5)

{

last;

}

8 BIT AVENUE

Perl Sort Hash by Value

More from my site

About Author

Mohammed Abualrob

Add a Comment