my $single;
sub new {
    return $single if defined $single;
    # set up args
    $self->{dbh} = DBI->connect(@args);
    $single = bless($self, $class);
    return $single;
}

# procedural 
{
    my $dbh;
    sub get_handle {
        return $dbh if defined $dbh;

        # handle arguments
        $dbh = DBI->connect(@args);
    }

}
[download]

try DBI->connect_cached. (like so)

use DBI;

...


$dbh=DBI->connect_cached( ... ) or die ...;
[download]

After this call your $dbh handle is live and connected and ready to use. If the handle is already connected it just uses that if not it reconnects for you.

Your parameters to the connect must be identical for this to reuse an existing connection (this is normally the case and shouldn't be a problem)

you could even put it in an eval to handle the die condition maybe sleeping and trying again depending on the specific error.

There is also prepare_cached etc.

Enjoy

die "no database handle" unless ($dbh && $dbh->ping)

-Blake

But if you have more than one potential data source (or if you like the visual effect ;) you can use it to wait and initialize the objects as needed.

#!/usr/bin/perl -w
use strict;

package Queries;

my ($db_handle, $ldap_connection);

sub data_query_one {
    $db_handle ||= new Sybase::DBlib();
    ...
}

sub data_query_two {
    $db_handle ||= new Sybase::DBlib();
    ...
}

sub ldap_query_one {
    $ldap_connection ||= new LDAP::Conn();
    ...
}

# &c.


sub clean_up {
    $ldap_connection->close     if $ldap_connection;
    $db_handle->release         if $db_handle;
}

1;
[download]

If a DBI->connect fails, there will be an error string set in DBI::errstr. This is analogous to the $dbh->errstr when you have an open db handle. The perldoc on DBI will have more info.

e.g.

my $dbh = DBI->connect('DBI:stack_of_papers_on_my_desk', 'user', 'pw')
    or die DBI::errstr;
[download]

elbieelbieelbie