The mann_whitney function performs a Mann-Whitney U-test on two random samples drawn with different queries. The mann_whitney function takes four parameters:
  1. the numeric field to perform the Mann-Whitney U-test on
  2. the Lucene/Solr query for drawing the first sample
  3. the Lucene/Solr query for drawing the second sample
  4. the sample size for both samples

​
Sample syntax

select mann_whitney(sepal_width_d, "species_s:versicolor", "species_s:virginica", 150) as u,
       p_value
from iris

​
Result set

The result set for the mann_whitney function contains one row with the results of the mann_whitney test. The mann_whitney function returns the U-statistic. The p_value field can be selected to return the p-value for the Mann-Whitney test. Sample result set in Apache Zeppelin Sample result

​
Visualization

Sample visualization of the mann_whitney function using the Apache Zeppelin number visualization: Sample visualization