The JDOQL
SELECT this.id FROM org.jpox.test.A
WHERE this.setB.contains(b1) && b1.propertyName == :propName1 && b1.value == :propValue1
VARIABLES org.jpox.test.B b1

results in SQL of

SELECT `THIS`.`ID`
FROM `A` `THIS` CROSS JOIN `B` `UNBOUND_B1`
WHERE EXISTS (
    SELECT 1 FROM `B` `THIS_SETB_B1`
    WHERE `THIS_SETB_B1`.`A_ID` = `THIS`.`ID`
    AND `THIS_SETB_B1`.`A_ID` = `THIS`.`ID`
    AND `UNBOUND_B1`.`ID` = `THIS_SETB_B1`.`ID`
    AND `THIS_SETB_B1`.`PROPERTY_NAME` = 'EditorialItemTitle'
    AND `THIS_SETB_B1`.`VALUE` = 'Plan the Perfect Alaskan Vacation')

so the EXISTS subselect has THIS_SETB_B1.A_ID = THIS.ID duplicated. This comes out of CollectionExpression.contains

QueryExpression qexpr = collStore.getExistsSubquery(qs, mapping, te, ctRangeVar);
...
ScalarExpression bindTo = collStore.joinElementsTo(qexpr, qs, mapping, te, ctRangeVar, var.getVariableType(), expr, expr.te == null ? etRangeVar : expr.te.getRangeVariable());

and each of these parts is adding the constraint. Why ?