| 1 |
root |
1.1 |
package Tree::M; |
| 2 |
|
|
|
| 3 |
|
|
use Carp; |
| 4 |
|
|
use DynaLoader; |
| 5 |
|
|
|
| 6 |
|
|
BEGIN { |
| 7 |
|
|
$VERSION = 0.01; |
| 8 |
|
|
@ISA = qw(DynaLoader); |
| 9 |
|
|
bootstrap Tree::M, $VERSION; |
| 10 |
|
|
} |
| 11 |
|
|
|
| 12 |
|
|
=head1 NAME |
| 13 |
|
|
|
| 14 |
|
|
Tree::M - implement M-trees for efficient "multimedia-searched" |
| 15 |
|
|
|
| 16 |
|
|
=head1 SYNOPSIS |
| 17 |
|
|
|
| 18 |
|
|
use Tree::M; |
| 19 |
|
|
|
| 20 |
|
|
=head1 DESCRIPTION |
| 21 |
|
|
|
| 22 |
|
|
(not yet) |
| 23 |
|
|
|
| 24 |
|
|
Ever had the problem of managing multi-dimensional (spatial) data but your |
| 25 |
|
|
database only had one-dimensional indices (b-tree etc.)? Queries like |
| 26 |
|
|
|
| 27 |
|
|
select data from table where latitude > 40 and latitude < 50 |
| 28 |
|
|
and longitude> 50 and longitude< 70; |
| 29 |
|
|
|
| 30 |
|
|
are quite inefficient, unless longitude and latitude are part of the same |
| 31 |
|
|
spatial index (e.g. a r-tree). |
| 32 |
|
|
|
| 33 |
|
|
An M-tree etc.. etc... |
| 34 |
|
|
|
| 35 |
|
|
=head2 THE Tree::M CLASS |
| 36 |
|
|
|
| 37 |
|
|
=over 4 |
| 38 |
|
|
|
| 39 |
|
|
=item $M = new Tree::M arg => value, ... |
| 40 |
|
|
|
| 41 |
|
|
... |
| 42 |
|
|
|
| 43 |
|
|
distance => specify a distance function. the default distance |
| 44 |
|
|
method uses this function in some way. |
| 45 |
|
|
keyof => the keyof method uses this argument to determine |
| 46 |
|
|
the key of an object |
| 47 |
|
|
min => the minimum number of objects stored in a node |
| 48 |
|
|
max => the maximum number of objects stored in a node |
| 49 |
|
|
|
| 50 |
|
|
=cut |
| 51 |
|
|
|
| 52 |
|
|
=back |
| 53 |
|
|
|
| 54 |
|
|
=head1 AUTHOR |
| 55 |
|
|
|
| 56 |
|
|
Marc Lehmann <pcg@goof.com>. |
| 57 |
|
|
|
| 58 |
|
|
=head1 SEE ALSO |
| 59 |
|
|
|
| 60 |
|
|
perl(1), L<DBIx::SpatialKeys>. |
| 61 |
|
|
|
| 62 |
|
|
=cut |
| 63 |
|
|
|
| 64 |
|
|
1; |
| 65 |
|
|
|
