Filter_Query ONLY stores the IDS of the document. This makes the filter very quick to include / exclude documents. Good examples of this are filtering products from a search based on country, product type, availability, etc.
A regular query can perform the same function, but it has a very sophisticated scoring system to determine "relevance." I believe that the documentation indicates that scoring is done only in the main request, and not in the filter. This should also increase the speed of the request.
So, I can request for:
description:Kohler AND productType:Toilet
Or I can request for:
description:Kohler with a FQ of productType:Toilet
The results will be the same, but the ratings will be different. In addition, if you receive many different requests during the day, which for productType:Toilet , FilterQuery will be cached, which will speed up the overall request time.
rfeak Jan 04 '12 at 17:20 2012-01-04 17:20
source share