Zero-count terms and minCount=0 might lead to duplicate entries

tokee / lucene-solr

High cardinality faceting (SOLR-5894)

http://tokee.github.io/lucene-solr/

7 stars 1 forks source link

Zero-count terms and minCount=0 might lead to duplicate entries #17

Closed tokee closed 9 years ago

tokee commented 9 years ago

The optimization of minCount=0 with sparse potentially gives double entries. This can be fixed by making the first (sparse) pass with minCount=1 and the second (sequential) pass with count=0.