lucy-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Thomas den Braber" <tho...@delos.nl>
Subject [lucy-user] Hits offset and search performarce
Date Sat, 10 Nov 2012 19:13:32 GMT
Hallo,

I am migrating from Swish-e to Lucy, most of the features I have implemented without many
troubles but I have found some (big) differences in performance when using the offset
parameter in the hits object.

If I do:

my $hits = $searcher->hits( 
	query => $queryobj,
	sort_spec => $sort_spec,
	num_wanted => 20,
	offset => 5000
);

The search time is much higher then with the offset set to 0. In Swish-e there is only a
small increase in search time.

I found out that when running with  num_wanted => 5000 and offset => 0 the search time
is
almost the same as with  num_wanted => 20 and offset => 5000. It looks like there is
no
optimization when using 'offset'. 
I would expect that using the offset, performance should be higher because no processing
needs to be done to the hits before the offset (no score calculation).

Is there a way to improve the search performance when using the 'offset' ?
Could this be something for future versions ? 

Also I am missing the seek command (that exists in swish-e) to fast forward through my
results or did I overlook something ?

But in general I am very satisfied with Lucy so far. Thanks for the work guys !

Thomas



Mime
View raw message