Re: sorting and merging in perl

[Veltro]

Hi Sekhar Reddy,

I think something like this can help you out:

use strict ;
use warnings ;
my %results = () ;
while ( <DATA> ) {
    chomp ;
    my @row = split /,/, $_ ;
    if ( exists $results{ $row[0] } ) {
        # Update existing hash entries
        if ( ( $row[2] ) <
             $results{ $row[0] }->{ 'MIN ACTDATE' }
        ) {
            $results{ $row[0] }->{ 'B1 MIN ACTDATE' } = $row[1] ;
            $results{ $row[0] }->{ 'MIN ACTDATE' } = $row[2] ;
        }
        if ( !( defined $row[3] &&
                defined $results{ $row[0] }->{ 'MAX DEACTDATE' } )
        ) {
            $results{ $row[0] }->{ 'B1 MAX ACTDATE' } = undef ;
            $results{ $row[0] }->{ 'MAX DEACTDATE' } = undef ;
        } elsif ( ( $row[3] ) >
                  ( $results{ $row[0] }->{ 'MAX DEACTDATE' } )
        ) {
            $results{ $row[0] }->{ 'B1 MAX ACTDATE' } = $row[1] ;
            $results{ $row[0] }->{ 'MAX DEACTDATE' } = $row[3] ;
        }
    } else {
        # Create new entry in hash
        $results{ $row[0] } = {
            'A1' => $row[0],
            'B1 MIN ACTDATE' => $row[1],
            'B1 MAX ACTDATE' => $row[1],
            'MIN ACTDATE' => $row[2],
            'MAX DEACTDATE' => $row[3],
        }
    }
}
foreach ( sort keys %results ) {
    my $a1 = $results{ $_ }->{ 'A1' } ;
    my $b1ma = $results{ $_ }->{ 'B1 MIN ACTDATE' } ;
    my $b1md = $results{ $_ }->{ 'B1 MAX ACTDATE' } // 'NULL' ;
    my $mad = $results{ $_ }->{ 'MIN ACTDATE' } ;
    my $mdad = $results{ $_ }->{ 'MAX DEACTDATE' } // 'NULL' ;
    print "$a1,$b1ma:$b1md,$mad,$mdad\n" ;
}
__DATA__
7900724666,200906888,20180416,20180522
7900724666,200906888,20180601,20180720
7900724666,200906888,20180406,20180411
7900724677,200906872,20180301,20180330
7900724677,200906871,20180101,20180228
7900724677,200906873,20180401,20180420
7900724688,200906881,20180101,20180228
7900724688,200906881,20180303,20180330
7900724688,200906882,20180404,20180430
7900724688,200906883,20180508,20180620
7900724699,200906891,20180101,20180228
7900724699,200906891,20180303,20180330
7900724699,200906892,20180404,20180430
7900724699,200906893,20180508,

Assuming the first columns is a unique key, I am using a hash to store the data for that unique key.

The output is:

7900724666,200906888:200906888,20180406,20180720
7900724677,200906871:200906873,20180101,20180420
7900724688,200906881:200906883,20180101,20180620
7900724699,200906891:NULL,20180101,NULL

With best regards,
Veltro

[Sekhar Reddy]

Thank you somuch Veltro, Greatly appreciated your help. Once again thank you