LP:686353 - 2.7x performance regression with correlated indexes and sort_intersection in maria-5.1-wl21

The following query against the DBT-3 dataset

SELECT count(*) FROM lineitem WHERE

( l_receiptDATE BETWEEN '1992-11-01' AND '1992-11-29' ) AND l_shipdate < '1992-11-15';

is 10x times slower when executed with sort_intersection.

Igor says that this is " not interesting because lineitem is naturally ordered by l_receiptDATE and l_receiptDATE is strongly correlated with l_shipdate" , " in the case when indexes are correlated intersection causes degradation: this is a well know fact.".

At the same time, the data set is supposedly standard and in real life, almost all date-based indexes will be naturally ordered and correlated to the PK.