I haven't tried this "workaround" yet, but until we have a Search API
available, it might work...
Try uploading your products to Google Base and then using their API to
search your own stuff. We upload all our books to google base, and if
I go to http://base.google.com/ and search for our domain
Yes, you're encountering 'exploding indexes'. The number of index
entries needed for a 3-term index is, for example, n*(n-1)*(n-2).
My suggestion would be this:
- Keep indexes for one and two uses of the term list.
- Keep a separate entity type that tracks how many occurrences there
are of each t
There's a definite need for a full-text search API (particularly
substring matches) that just works for the datastore.
For public-facing information like your products, you might consider
outsourcing the search to something like Google Custom Search:
http://www.google.com/coop/cse/
(There's an ad
Why not just do a single word query, then do a brutal force scan on
result?
On Dec 7, 4:37 pm, Koen Bok <[EMAIL PROTECTED]> wrote:
> First off, I _really_ like the idea of AppEngine. We're hosting most
> of our sites (versionsapp.com, checkoutapp.com) on it and it works
> great for that purpose.
Hmmm, your right.
You could try culling some of the token words, the simple
ones like 'and', 'a', 'or' to limit the number of puts,
but that will only go so far.
I don't think sorting and checking for intersection will
cost you too much. Its the number of results returned from
the queries that
Hey Lock,
I thought about your idea before, but it also has huge downsides:
- Inserting a single entity with 40 token words (not uncommon) will
lead to 41 inserts, and inserting is BigTables biggest weakness. Same
goes for deleting.
- It's (I think) not possible with this setup to get the second
I'm pretty new to app engine so there may be some
gotcha's in what I say, but I think my logic is ok.
It seems you have a pretty simple table with a few
fields (name, brand, description) and you want to be
able to search any one of the words in any of the
fields.
I'd suggest building a search te