Starting from 1.4, solr handles faces with a large number of values ​​perfectly, since by default a simple number of facets is used. (by default, facet.method is "fc").
Prior to 1.4, solr used a filter-based facet method (enum), which is certainly faster for cutting an attribute with few values. This method requires one filter per facet.
About facet.limit, think of it as a way to navigate the facet space (in conjunction with facet.offset), for example, you navigate through the resulting space with lines / offset. Therefore, a value of 10 ~ 50 is reasonable.
As with lines / offsets, and due to the nature of Solr, you can expect that the performance of facet.limit / facet.offset will degrade when the offset gets larger, but should be fine if you stay within reasonable limits.
By default, solr produces more frequent faces first.
Summarizing:
Use Solr 1.4
Make sure facet.method is "fc" (well, by default, anyway).
Go through the facet space with facet.limit / facet.offset.
jeje
source share