I'm not sure about what you want to do, but I hope this helps:

#!perl
use strict;
use warnings;
use Data::Dumper;

my $id = '1C8962EB';
myfunction($id, { max => 90, cval => 89 });
sub myfunction {
    my %p = %{$_[1]};
    print Dumper(\%p);
    my %newp = map { $_ => { value => $p{$_}, key => $_, id => $_[0] }
+ } keys %p;
    print Dumper(\%newp);
}
__END__
$VAR1 = {
          'max' => 90,
          'cval' => 89
        };
$VAR1 = {
          'cval' => {
                      'value' => 89,
                      'id' => '1C8962EB',
                      'key' => 'cval'
                    },
          'max' => {
                     'value' => 90,
                     'id' => '1C8962EB',
                     'key' => 'max'
                   }
        };
[download]

Thanks it does help, I was just wondering if there was any way to avoid having to use map or another transformation at all.

To clarify the original objective, it was to have a common hash structure between the function I mentioned and results from selectall_hashref.

I was trying to find a way to manipulate the selectall_hashref to provide a resulting hash as described that can be used as $$hash{$id}{$key} = $value.

Your code provides a valid solution albeit through a different approach - "modify the structure passed to the function to look like the result from the selectall hash", and should work just fine unless someone can come up with an even better solution.

If I understand the question (and your subsequent clarification) correctly, I think this should do the trick:

use strict;
use warnings;

use DBI;
use DBD::SQLite;
use Data::Dumper;

my $dbh = DBI->connect(
    'dbi:SQLite:dbname=test.sqlite',
    undef, undef, { RaiseError => 1, PrintError => 1, } );

eval {
    $dbh->do('CREATE TABLE evkeypairs (id INTEGER PRIMARY KEY NOT NULL
+,key TEXT,value TEXT);');
    $dbh->do(q!INSERT INTO evkeypairs (key, value) VALUES ('foo','bar'
+);!);
    $dbh->do(q!INSERT INTO evkeypairs (key, value) VALUES ('kung','foo
+');!);
} or warn "$@";

my $sth = $dbh->prepare( 'SELECT * FROM evkeypairs' );
$sth->execute;
my $href = $sth->fetchall_hashref( 'id' );

print Dumper $href;

$dbh->disconnect;
unlink 'test.sqlite';
__END__
$VAR1 = {
          '1' => {
                   'value' => 'bar',
                   'id' => '1',
                   'key' => 'foo'
                 },
          '2' => {
                   'value' => 'foo',
                   'id' => '2',
                   'key' => 'kung'
                 }
        };
[download]

Thanks to AM for providing the code to create a test db.

#!/usr/bin/perl -- 

use strict;
use warnings;

use DBI;
use DBD::SQLite;
use Data::Dumper;

{
    my $dbh = DBI->connect(
        'dbi:SQLite:dbname=test.sqlite',
        undef, undef, { RaiseError => 1, PrintError => 1, },
    );

    eval {
        $dbh->do('CREATE TABLE evkeypairs (id INTEGER PRIMARY KEY NOT 
+NULL,key TEXT,value TEXT);');
        $dbh->do(q!INSERT INTO evkeypairs (key, value) VALUES ('foo','
+bar');!);
        $dbh->do(q!INSERT INTO evkeypairs (key, value) VALUES ('kung',
+'foo');!);
    } or warn "$@";

    print Dumper( $dbh->selectall_hashref('select * from evkeypairs ',
+ [ 'id','key']) ),"\n";
    print Dumper( $dbh->selectall_hashref('select * from evkeypairs ',
+ 'id') ),"\n";

    $dbh->disconnect;
    use autodie qw' unlink ';
    unlink 'test.sqlite';
}
__END__
$VAR1 = {
          '1' => {
                   'foo' => {
                              'value' => 'bar',
                              'id' => '1',
                              'key' => 'foo'
                            }
                 },
          '2' => {
                   'kung' => {
                               'value' => 'foo',
                               'id' => '2',
                               'key' => 'kung'
                             }
                 }
        };

$VAR1 = {
          '1' => {
                   'value' => 'bar',
                   'id' => '1',
                   'key' => 'foo'
                 },
          '2' => {
                   'value' => 'foo',
                   'id' => '2',
                   'key' => 'kung'
                 }
        };
[download]

I am actually trying to avoid having the above hash structure in the related function I mentioned, since this function deals with many different attributes that I find convenient to provide via the hash.

If I were to use a similar hash structure to your example in this function, it would now need to be called in the following (tedious) way:

myfunction($id,{ kung => { value => foo }, foo => { value => bar}});
[download]

Here is the ideal hash structure with some example attributes that I am trying to attain through the select:

Attributes now: $VAR1 = {
          '49CBAF4B' => {
                          'occurrences' => 999,
                          'object' => 'device123',
                          'timestamp' => 1261660376,
                          'value' => 90,
                          'max' => 96
                        }
        };
[download]

If this cannot be done, I might be forced to map the results to a hash with the desired layout, or look at a different approach altogether.

Niel