Re: Nested Sets vs. Nested Intervals

asdf wrote:
> Thank you very much, Vadim Tropashko.
>
> Nested Intervals are very efficient for updating and selecting
> hierarchical categories. I am creating a web application/directory from
> it, but it would be more convenient for the user to retrieve its
> children from an array of folder names in the URL (seperated by dots or
> slashes, like google groups does for its usenet categories) instead of
> a category_id.
>
> Both dmoz and yahoo also retrieves its immediate subcategories by the
> array of folder-names in the URL; Unlike some other directories that
> retrives its immediate subcategories from a category_id in the URL. Is
> there any efficient way to retrieve its immediate subcategories by an
> array of folder-names such as dmoz and yahoo?
>
> Basically, it's like replicating a file system: an array of folder
> names can point to the category_id to retrieve their children. However,
> it would take recursive SELECTs to find the category_id from the array
> of folder names. I could store the whole path in each row, but it would
> be very inefficient to update a category. Is there an efficient way to
> do this?

I assume that you asked how to query immediate descendants (children!) in the nested intervals model. I've emailed the chapter snippet on nested interval encoding to your yahoo email address. Look up page 15 where it develops that nested intervals matrix encoding is an adjacency model. Received on Thu Nov 10 2005 - 01:00:22 CET