first experiment to transfer data from Reblog to MongoDB

diff --git a/reblog2mongodb.pl b/reblog2mongodb.pl
new file mode 100755 (executable)
index 0000000..8ecb90a
--- /dev/null
+++ b/reblog2mongodb.pl
@@ -0,0 +1,109 @@
+#!/usr/bin/perl -w
+
+use strict;
+use DBI;
+use MongoDB;
+use Data::Dump qw/dump/;
+
+$|++;
+
+my $debug = @ARGV ? 1 : 0;
+
+my $database = 'reblog';
+
+my $dbi = "DBI:mysql:database=$database";
+$dbi .= ";host=127.0.0.1;port=13306";  # XXX over ssh
+
+my $dbh = DBI->connect($dbi,"","",{ RaiseError => 1 });
+
+$dbh->do(qq{
+       create temporary table published_items as
+       select
+               item_id
+       from
+               items_userdata
+       where
+               label = 'published' and
+               value_numeric = 1
+});
+
+my $sql = qq{
+       select
+               i.id as item_id,
+--             i.guid as _id,
+--             i.link as _id,
+               i.*,
+               f.url as feed_url,
+               f.title as feed_title,
+               f.link as feed_link,
+               f.description as feed_description
+       from items i
+       join published_items p on i.id = p.item_id
+       join feeds f on i.feed_id = f.id
+       where i.id > ?
+       order by i.id asc
+       limit 1000
+};
+
+my $sql_tags = qq{
+select
+       items_userdata.item_id,
+       value_long as tags,
+       timestamp
+from items_userdata
+join published_items p
+       on items_userdata.item_id = p.item_id and label='tags'
+where
+       items_userdata.item_id > ?
+order by items_userdata.item_id asc
+};
+
+my $conn  = MongoDB::Connection->new;
+my $db    = $conn->get_database( $database );
+my $items = $db->get_collection( 'items' );
+
+my $last_row = 0;
+$last_row = 0 if $debug;
+
+print "Fetching items from $dbi id > $last_row\n";
+
+my $sth = $dbh->prepare($sql);
+$sth->execute( $last_row );
+
+warn dump( $sth->{NAME} );
+
+print "found ",$sth->rows," items to process...\n";
+
+my $sth_tags = $dbh->prepare($sql_tags);
+$sth_tags->execute( $last_row );
+print "found ",$sth_tags->rows, " tags found...\n";
+
+my $count = 0;
+
+my $row_tags = $sth_tags->fetchrow_hashref();
+
+while (my $row = $sth->fetchrow_hashref() ) {
+       my $_id = $row->{_id} || "c$count";
+       $_id =~ s{\W+}{_}g;
+       $_id =~ s{_+$}{};
+
+       my $doc = $row;
+#      $row->{_id} = $id;
+
+       while ( $row_tags && $row_tags->{item_id} < $row->{item_id} ) {
+               $row_tags = $sth_tags->fetchrow_hashref();
+               warn "## got tags: ",dump( $row_tags ) if $debug;
+       }
+
+       if ( $row_tags && $row_tags->{item_id} == $row->{item_id} ) {
+               $doc->{tags} = [ split(/\s+/, $row_tags->{tags} ) ];
+               warn "++ ",$row->{item_id}, dump( $row->{tags} ),$/;
+       }
+
+       $items->insert( $doc );
+
+       $last_row = $row->{id};
+       $count++;
+
+}
+