lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dawid Weiss <>
Subject Re: Storing and loading the FST directly from disk
Date Sat, 04 Jun 2011 08:54:41 GMT
> Wow you're right, the FST size in RAM with 50 mil date 1 ms
> incremented keys is less than 1K.  That's insane!

This does sound insane. Are you sure you're building everything right
(not pruning anything)? You could always enumerate the FST to get the
keys back to make sure it's actually working. Or check for exist(key)
for every key in the input.

The compression ratio should be good for shared prefixes, but 1K seems
a bit too small for 50mil entries...


To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message