Hi, Need your expert advice on fine tuning the below perl script, which is currently taking more than 30 minutes to extract information from a raw file which has nearly 2million records. below are the requirements met in the script: 1) script should scan through a log file and output the final result into a file called "test.txt" 2)it replaces some strings in the log file into more generic terms 3)it finds the record number of the affected row from the log file and uses this record number to query the main source file for the particular record 4) it generates a string which now contains information from the logfile and patches it along with the affected row from the source 5) the result is written into the output file test .txt

=======================code ============================
#!/usr/bin/perl

$read_file = "$ARGV[0]";

$read_source = "$ARGV[1]";

open(LOGFILE,$read_file) or die "An Error Occured : $!";

open(REPORT,">/retsit/systematics/test.txt");

$str1 = 'failed all WHEN clauses';

$str2= 'CUST SEGEMENT IS EMPTY';

$str3= 'unique constraint';

$str4= 'DUPLICATE RECORD';

  while(<LOGFILE>)

     {

       if ($_ =~ /Record/)

           { $_  =~  s/$str1/$str2/g;

             $_ =~ s/$str3/$str4/g;        

             $ind1 = index($_,'Record')+6;

             $len2 =index($_, ':')-6;

             $recnum = substr($_,$ind1,$len2);

             $recnum =~ s/^\s+|\s+$//g ;

            $strx = "sed -n '".$recnum."p' ".$read_source;

            $str5 = `$strx`; 

            $_ .= '|'."$recnum"."$str5\n";

            print REPORT $_; 

           }

     }

  close(LOGFILE); 

  close(REPORT);
[download]