On Thu, Nov 20, 2014 at 1:08 PM, Gustavo Duarte <gus.dua...@gmail.com> wrote:
> My doubt is, when I make a movie search by title in a another different
> language than English, why the search is more precise if use HTTP access
> than SQL ?

They just have a different set of information.
The plain text data files (used by SQL) is somehow limited; on the
IMDb many more information came from third-party companies
and are licensed in a way that prevents them to be included in
the plain text data files.

In your case, if you open the aka-titles.list file, you will see that
"El gran robo"
is not associated to Flypaper at all (it's, instead, correctly
associated to 7 uomini e un cervello, but that's obviously
another movie).

So, it has no way to know that your El gran robo is in fact

Plus, even having all the data, I do a limited kind of search on
titles and AKAs:
they are first matched (with/without articles like The, A, An, ...) to
the entries
in the database using a soundex algorithm (to match things that
"sounds similar"),
then the matching subset is compared and sorted using a distance algorithm.
It works quite well, in fact, but on the IMDb site they can use much more
sophysticated algorithms and also meta-information like how many people
are looking for a given title, how much popular a movie is and so on.

So, in your specific case: no way.  In general: it's not that bad, but
I'm obviously
open to receive patches that improve the search. :-)

Davide Alberani <davide.alber...@gmail.com>  [PGP KeyID: 0x465BFD47]

Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
from Actuate! Instantly Supercharge Your Business Reports and Dashboards
with Interactivity, Sharing, Native Excel Exports, App Integration & more
Get technology previously reserved for billion-dollar corporations, FREE
Imdbpy-help mailing list

Reply via email to