Hello, my name is Nir and I’m a big fan of Lucene. Lately I have investigated the inner classes of Lucene to understand things like the order a conjunction query is executed.
A strange situation happened to me when I compose a Boolean query with a term query and a prefix query. I found that the cost of the prefix query was much lower than the cost of the term query although it has much more docs associated with it. I tried to understand the calculation of the cost of a prefix query but found it unclear. (counter / termsPerDoc). I would be very happy if someone would be able to explain me the calculation. In Addition I would like to make sure I understand the cost() function correctly. Isn’t it suppose to help the iterators of a query iterate the least they can? Thanks in advance
