I would go farther in use of the database, actually. Consider "I get 100s of images every month that needs to be uploaded to the server." The approach seems to be to just toss them in piles on the floor, with "the image path and file name from the DB table". Boo.

I like storing the actual images in a DB. Makes it easier to manipulate, record attributes, etc. For example, in a current project, I'm storing a bunch of process maps. The maps correspond to annotated visio diagrams. We trot them out, tart them up based on state the users have reported, and then let the users get a visual image on how their process is going.

The visio's get converted into PNG's, and dumped into a file directory on someone's local drive. They are then read into the SQL server for general use. Here is the sample code that stores the images into the database:

my $sth = $dbh->prepare("SELECT processID, processCode FROM processLis
+t") or die "bad db: ".$dbh->errstr();
$sth->execute or die "worse DB: ".$dbh->errstr();

my $sth2 = $dbh->prepare("UPDATE blobs SET graphic = ? WHERE processID
+ = ?") or die "ugly DB: ".$dbh->errstr();
while (my $ref = $sth->fetchrow_hashref) {
  my $processID   = $ref->{'processID'};
  my $processCode = $ref->{'processCode'};
  my @names       = glob($processCode . "_*.PNG");
  if ($#names != 0) {
    die "problem with $processCode, got ", join(", ", @names);
  }
  my $image = Image::Magick->new;
  my $rv    = $image->Read($names[0]);
  die $rv if $rv;
  my $blob = $image->ImageToBlob();
  $sth2->execute($blob, $processID) or die "can't store blob for $name
+s[0] under processID $processID: $!";
  print "Stored blob $names[0] for processID $processID\n";
}
[download]

--woody