Implement Efficient Personalized Search

["刁轶夫 (Yifu Diao)" <al...@gmail.com>]

Hi,
I am using Django-Haystack which connects Django with Solr.
I have a model called Item which is about webpages. I want to enable
personalized search so that users can search only the items they have
shared.
I solved the problem like this:

sqs = searchqueryset.auto_query(query).filter(primary_key__in=own_items_ids)
own_item_ids is a list of integers that I got from Django ORM.
primary_key is just the primary key of the Item model.

My questions is,
By observing the actual query in Solr, there are lots of OR(like
item_id=1 OR item_id=2 ...) So I guess the time complexity is O(n^2),
but I want to reduce it to O(n). I want to build a hash table of
own_item_ids, and query this hash table while scanning the results
linearly so that I can get the filtered results.
Is this possible, if yes, how?
I have passed this question to Haystack maillist, but the author said
he doesn't know about the underlying mechanism.

Thanks!
Yifu