optimization - Lucene.Net memory consumption and slow search when too many clauses used -

itemprop = "text">

I have a DB that is the text file attributes and text file primary key ID and about 1 million text files with their ID Indexed (primary key in DB)

Now, I am searching on two levels. Firstly, directly DB search, where I get the primary key as a result (approximately 2 or 3 million id)

Then I create a boolean query for example as the following

< P> + text: "test *" + (PKID: 1PKID: 4PKID: 100PKID: 115PKID: 1041 ....)

and in my index file Find it.

The problem is that such questions (containing 2 million volumes) take too much time to give results and eat too much memory ....

Is there any optimization solution for this problem?

Assume that you can reuse the dbid part of your questions: < / P>

1. Split query in two parts: a part (text query) query will be created and the second part (PKID query) filter will be created
2. Make both parts in the query
3. Convert a pkid query for filter (using QueryWrapperFilter)
4. Convert filters to a cached filter (using caching spoiler filter)
5. F Wait on the filter, sharpen
6. Next time you search, use surcharges that allow you to use a query and filter

until To reuse pkid search, you should make major improvements. As long as you do not optimize your index, the effect of caching should also work through committed digits (I think the bit set is calculated on a per-segment basis).

HTH

<

ps

I think it will not be remembered that I think that you can use it to abuse your point of view Are putting!