lucidworks/blog-sortminmax-perf

Name: blog-sortminmax-perf

Owner: Lucidworks

Description: Performance comparison of sorting on single valued (long) fields compared to sorting on the 2 arg "field" function added in SOLR-2522

Created: 2015-09-09 23:16:56.0

Updated: 2017-10-06 07:57:17.0

Pushed: 2015-09-09 23:18:26.0

Homepage: null

Size: 136

Language: Perl

GitHub Committers

UserMost Recent Commit# Commits

Other Committers

UserEmailMost Recent Commit# Commits

README

blog-sortminmax-perf

Performance comparison of sorting on single valued (long) fields compared to sorting on the 2 arg “field” function added in SOLR-2522 (with the bug fixes applied in SOLR-8001).

Full Writeup at lucidworks.com

The random sample data generated be the included scripts indexes documents with multiple values in a “multi_l” field, and relies on updated processors to populate the corresponding “min_l” and “max_l” fields.

Note that in addition to comparing sort options such as min_l asc vs field(multi_l,min) asc, comparisons cab also be done of things like sum(min_l,max_l) asc vs sum(def(field(multi_l,min),0),def(field(multi_l,max),0)) asc (where the def() function is needed to ensure a comparable results due to the unresolved issue in SOLR-8005 since the random data includes roughly 10% of docs having no values in the specified fields.)


notes.txt shows the steps taken in creation of the data and running the various tests. It should include enough information to reproduce.


This work is supported by the National Institutes of Health's National Center for Advancing Translational Sciences, Grant Number U24TR002306. This work is solely the responsibility of the creators and does not necessarily represent the official views of the National Institutes of Health.