Find maximum number in text file and copying its full statement in new text file?

[sumathigokul]

Hi all,

i have a text file that contains the following lines

High fanout nets in the post compile netlist:
    Fanout  Type          Name
    --------------------------
    2       INT_NET       Net   : c_c
                          Driver: c_pad
    2       INT_NET       Net   : b_c
                          Driver: b_pad
    2       INT_NET       Net   : a_c
                          Driver: a_pad
    1       INT_NET       Net   : sum_c
                          Driver: sum_1_SUM0_0
    1       INT_NET       Net   : N_5
                          Driver: sum_1_CO0_i

I have to write a perl script that saves the following statements only, i.e. high fanout nets alone. For that, if i write a script which find maximum number in my text file and copies those statements in new file, it will copy the following statements.

    2       INT_NET       Net   : c_c
    2       INT_NET       Net   : b_c
    2       INT_NET       Net   : a_c

Here is the code i tried, but could not get the text file as i wish....

my $high_number = 0;
open (F1, "<new.txt") or die "cant open it$!";
while (<F1>) {
my ($number) = m/:(\d+):/;
$high_number = $number if $number > $high_number;
 }
print "Highest number found = $high_number\n";
close (F1);

please help me out in this, thank you all....

[Athanasius]

Hello sumathigokul,

Your regex m/:(\d+):/ finds a sequence of one or more digits occuring anywhere in the line between two colon characters. But in the text file shown, the only colons occur later in the line. You need something like this:

while (<F1>)
{
    my ($number) = /^\s*(\d+)/;

    if (defined $number)
    {
        $high_number = $number if $number > $high_number;
    }
}

in which the regex finds a sequence of digits at the beginning of the line, preceeded by zero or more whitespace characters only. Note also that if a match fails (as it does for lines 1, 2, 3, 5, 7, etc.), the variable $number is undefined. Always begin your scripts with:

use strict;
use warnings;

In this case warnings will alert you to the fact that $number is being used in a numerical comparison when it is undef.

Hope that helps,

[kcott]

G'day sumathigokul,

Your regex baffles me! You're trying to match (and capture) one or more digits immediately preceded by and followed with a colon — clearly not what you want. I'd recommend spending some time with the Perl Regular Expressions Tutorial.

You have no code to output the wanted lines — which is why that part isn't working. Take a look at open; note how both of the initial examples (for reading and writing) use the 3-argument form and lexical filehandles; start doing this yourself.

This code has the logic you need — you'll need to add the I/O:

#!/usr/bin/env perl

use strict;
use warnings;

my $re = qr{ \A \s+ ( \d+ ) }x;
my $high = 0;
my @lines;

while (<DATA>) {
    /$re/ or next;
    my $num = $1;
    next if $num < $high;

    if ($num > $high) {
        $high = $num;
        @lines = ();
    }

    push @lines, $_;
}

print "Highest number: $high\n";
print @lines;


__DATA__
High fanout nets in the post compile netlist:
    Fanout  Type          Name
    --------------------------
    2       INT_NET       Net   : c_c
                          Driver: c_pad
    2       INT_NET       Net   : b_c
                          Driver: b_pad
    2       INT_NET       Net   : a_c
                          Driver: a_pad
    1       INT_NET       Net   : sum_c
                          Driver: sum_1_SUM0_0
    1       INT_NET       Net   : N_5
                          Driver: sum_1_CO0_i

Output:

Highest number: 2
    2       INT_NET       Net   : c_c
    2       INT_NET       Net   : b_c
    2       INT_NET       Net   : a_c

[sumathigokul]

Hi kcott,

First of all thank you for your reply and Good day to you...I have modified the code, so that it extracts the data from one file and copies those statements to new file...But, its not copying...

 
use strict;
use warnings;

my $re = qr{ \A \s+ ( \d+ ) }x;
my $high = 0;
my @lines;
open (F1, "<designer.log") or die ;
open (F2, ">>new_lines.txt") or die;
while (<F1>) {
    /$re/ or next;
    my $num = $1;
    next if $num < $high;

    if ($num > $high) {
        $high = $num;
        @lines = ();
    }
    
    push @lines, $_;
    print F2 @lines;
}

print "Highest number: $high\n";

close (F1);
close (F2);

Correct me if its wrong....

[Anonymous Monk]

If you are going to print inside the loop, then use of @lines is useless (given the code), for the current line would be the only one (1) element of @lines, ever.

What is your input? What is the highest number printed?

[aaron_baugher]

This loops through the lines, getting the Fanout value from the lines with INT_NET, and if it's equal to or greater than the maximum Fanout value so far, it saves that line along with its Fanout value as a 2-element array reference in the array @lines.

Once it's done, it has the largest Fanout value in $max and all the lines with that value (along with some other lines, possibly) as the first elements in the array @lines. The map/grep statement goes through those, grepping for the ones with the $max Fanout value and then using map to pull out the line itself and pass that to print.

#!/usr/bin/env perl
use 5.010; use warnings; use strict;

my $max = 0;
my @lines;  # 2D array, each element holding a line and its Fanout val
+ue

while (<DATA>){
    if (/(\d+)\s+INT_NET/ and $1 >= $max ){
        push @lines, [$_,$1];
        $max = $1;
    }
}
print for map { $_->[0] } grep { $_->[1] == $max } @lines;

__DATA__
    Fanout  Type          Name
    --------------------------
    2       INT_NET       Net   : c_c
                          Driver: c_pad
    2       INT_NET       Net   : b_c
                          Driver: b_pad
    2       INT_NET       Net   : a_c
                          Driver: a_pad
    1       INT_NET       Net   : sum_c
                          Driver: sum_1_SUM0_0
    1       INT_NET       Net   : N_5
                          Driver: sum_1_CO0_i

[jeffa]

This "way to do it" uses a hash to find the highest number:

use strict;
use warnings;

my %candidates;
while (<DATA>) {
    push @{$candidates{$1}}, $_ if /^\s+(\d+)/;
}
my ($highest) = sort {$b<=>$a} keys %candidates;
print for @{ $candidates{$highest} };

__DATA__
High fanout nets in the post compile netlist:
    Fanout  Type          Name
    --------------------------
    2       INT_NET       Net   : c_c
                          Driver: c_pad
    2       INT_NET       Net   : b_c
                          Driver: b_pad
    2       INT_NET       Net   : a_c
                          Driver: a_pad
    1       INT_NET       Net   : sum_c
                          Driver: sum_1_SUM0_0
    1       INT_NET       Net   : N_5
                          Driver: sum_1_CO0_i