Andrzej Bialecki wrote:
Shouldn't this be combined with a HitCollector that collects only the
first-n matches? Otherwise we still need to scan the whole posting list...
Yes. I was just posting the work-in-progress.
We will also need to estimate the total number of matches by
extrapolating linearly from the maximum doc id processed. Finally, it
is probably rather slow for large indexes, whose .fdt won't fit in
memory. A simple way to improve that might be to use
Similarity.floatToByte-encoded floats when sorting (e.g., the norm from
an untokenized field) so that documents whose boosts are close are not
re-ordered. I'll start work on these in the morning. (It is currently
my middle-of-night.)
Doug
