Spider: query with long IN clause has performance issues

A simple SELECT query from SPIDER table is slow (~90sec) if WHERE has an IN-list with thousands (5558) of elements. InnoDB has no problems with this - the same query returns result in less than 1 sec.

If the number of elements is reduced, then SPIDER works well too - for example with 8 it takes .003 sec.

SELECT cause_id, location_id, year_id, sex_id, age_group_id, val AS current

FROM output_epi_single_year_v19063_prevalence

WHERE metric_id = 1

AND measure_id = 5

AND year_id = 2022

AND sex_id IN (2)

AND location_id IN (1,4,5,6,7,8,9,10)

AND age_group_id IN (3)

AND cause_id IN (294);

8 rows in set (0.003 sec)

DDL and EXPLAIN FORMAT=JSON added.