Faceting with Lucene Block Join Query - Lucene/Solr Revolution 2014
Download as PPTX, PDF1 like2,483 views
The document discusses the necessity of specialized faceting for Lucene block join queries, detailing the challenges with false positives when aggregating facet counts for products instead of SKUs. It introduces the proposed block join facet component and its expected functionality, emphasizing that facet counts should correspond to parent documents. The document concludes with suggestions for performance improvements and references for further reading.
Faceting with Lucene Block Join Query - Lucene/Solr Revolution 2014
- 2. Faceting with Lucene Block Join Query
- 3. Agenda 1. Why we need special faceting for Block Join queries? 1. Proposed Block Join facet component.
- 4. Introducing myself PRIVILEGED AND CONFIDENTIAL Oleg Savrasov, PhD A programmer Working for Grid Dynamics (griddynamics.com) Work and live in Saint-Petersburg, Russia
- 6. Jerrica is looking for a dress
- 7. Huge amount of dresses
- 8. Facet filters help Facet filters Reduced amount
- 9. Tasks to be solved ● Performant Search ● Facet calculation/filtering FacetComponent ?
- 10. Product has many SKU
- 11. Aggregated facet counts Facets should count products, not SKU. Expected facets: COLOR Blue : 1 Red : 1 SIZE S : 1 M : 1
- 12. Flat documents don’t help False positive match for +COLOR:Blue +SIZE:M
- 13. Separate SKU documents q = *:* facet.field = COLOR facet.field = SIZE COLOR Blue : 1 Red : 2 SIZE S : 2 M : 1 Wrong numbers! There is only one product
- 14. Search products only q = *:* fq = scope:product facet.field = COLOR facet.field = SIZE COLOR : 0 SIZE : 0 No such fields in product documents
- 15. Aggregated facet counts Facets should count products, not SKU. Expected facets: COLOR Blue : 1 Red : 1 SIZE S : 1 M : 1
- 16. Solr Block Join Support (since Lucene 3.4.0) Green Blue Yello w Yello w Blue Green Product Green Yello w Product Green Blue Yello w Yello w Product docId Child docs Parent doc Query: {!parent which="scope:product"}COLOR:Blue 1 1 1 Block 1 1 scope:product COLOR:Blue 1 ToParentQuery 1 1
- 17. SOLR-5743 Faceting with Block Join support ● Create BlockJoinFacetComponent ● ToParentQuery is expected ● Facet counts should correspond to amount of parent documents ● Only DocValues fields are supported
- 18. Faceting over DocSet slices Green Blue Yello w Yello w Blue Green Product Green Yello w Product Green Blue Yello w Yello w Product docId 0 1 0 0 1 0 1 DocSet Slice DocSet Slice counts COLOR Blue : 2 Aggregated counts COLOR Blue : +1
- 19. Block Join Facet Component
- 21. Facets counting
- 22. It works! q = {!parent which="scope:product"}COLOR:Blue child.facet.field = SIZE <response> ... <lst name="facet_counts"> <lst name="facet_fields"> <lst name="SIZE"> <int name="S">14</int> <int name="L">22</int> <int name="XL">17</int> </lst> </lst> </lst> </response>
- 23. The dress is found
- 24. Further improvements ● Thorough profiling ● Performance improvements ● Algorithmic improvements
- 25. References http://www.slideshare.net/MarkHarwood/proposal-for-nested-document-support-in- lucene http://blog.mikemccandless.com/2012/01/searching-relational-content-with.html http://blog.griddynamics.com/2013/09/solr-block-join-support.html
- 26. Big thanks! Please vote for SOLR-5743. Do you have any questions?