lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From dr <>
Subject some doubts about nodeblock in *.tim
Date Tue, 26 Jan 2016 10:05:12 GMT
Dear all ,
Currently, I'm studying  the format of *.tip and *.tim. I found that, in the *.tim it not
only stores suffixes information, but it also stores information of the  current prefix's
child nodes' position.Taks terms 'ab1,abc1,abc2‘ for example, for the prefix 'ab', *.tim
stores the suffix '1' and the file position of terms starting with prefix 'abc'. I wonder
why, because the file position of prefix 'abc' can be got form FST,  and prefix 'ab' knows
the existence of 'abc',why we store it in the *.tim again? For the speedup of WildCardQuery?
The version is 4.10.3.


  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message