Some aggregate functions, like COUNT(), SUM() or AVG() are order independent. But some are dependent, like: FIRST(), LAST(), STRING_AGG() etc. There are 2 different syntax types in SQL to provide ordering to the aggregate, within an aggregate:

SELECT string_agg(a, ',' ORDER BY a) FROM table;

OR using WITHIN GROUP construct:

SELECT LISTAGG(last_name, '; ') WITHIN GROUP (ORDER BY hire_date) "Employees";

FluentJPA supports both variants and you should refer to your vendor docs to choose the right one. The same in FluentJPA:

// first option with ALL or DISTINCT
SELECT(STRING_AGG(ALL(x.getA(), ",").ORDER(BY(x.getA()))));
SELECT(STRING_AGG(DISTINCT(x.getA(), ",").ORDER(BY(x.getA()))));

// second option
SELECT(alias(aggregateBy(LISTAGG(emp.getLastName(), "; "))
                        .WITHIN_GROUP(ORDER(BY(em.getHireDate()))), "Employees"));