You are viewing a plain text version of this content. The canonical link for it is here.
Posted to java-user@lucene.apache.org by 柯伟 <kk...@163.com> on 2020/02/14 20:04:56 UTC
create two level of index in CompressingTermVectorsWriter is a bit
redundant
In Lucene80DocValuesConsumer, we use `addTermsDict()` to create the first index level, and use `writeTermsIndex()` to create the second index level, terms dict needs to be traversed in each function, which will be very costly. We can obtain the two index levels only by traversing once.
If we could reduce one traversal?