JPA Create Criteria with GROUP COUNT and HAVING on nested objects - jpa

Question

Given this SQL query
SELECT
ug.lookup_key,
count(ug.id) as count
FROM user u
INNER JOIN user_group ug on ug.id = u.id
WHERE
u.age >= 11 AND
u.age <= 20 AND
ug.lookup_key in('12345')
GROUP BY ug.lookup_key
HAVING count(ug.id) < 7
I have written this
CriteriaBuilder criteriaBuilder = entityManager.getCriteriaBuilder();
CriteriaQuery<Object[]> query = criteriaBuilder.createQuery(Object[].class);
Root<UserGroup> d = query.from(UserGroup.class);
Join<UserGroup, User> join = d.join("users");
Predicate pred1 = criteriaBuilder.between(join.get("age"), ageFrom, ageTo);
Expression<String> exp = d.get("lookupKey");
Predicate pred2 = exp.in(lookupKeys);
query.where(pred1, pred2);
query.multiselect(d.get("lookupKey"), criteriaBuilder.count(d.get("id"))).groupBy(d.get("lookupKey"));
List<Object[]> results = entityManager.createQuery(query).getResultList();
for(Object[] object : results){
System.out.println(object[0] + " " + object[1]);
}
The SQL returns {"12345",4} whereas the code returns {"12345", 37}. The SQL is the correct result. There are 37 users in the database for groups with that lookup key, so I understand where the numbers are coming from but I do not understand how to do the JOIN, GROUP BY, HAVING with the CreateCriteria query so that I get the results. I don't want to use JPQL.
The entities...
#Entity
public class User {
#Id
#GeneratedValue(strategy = GenerationType.AUTO)
private long id;
private String name;
private int age;
private double salary;
#ManyToOne(optional=false,cascade=CascadeType.ALL, targetEntity=UserGroup.class)
#JsonBackReference
private UserGroup group;
// Getters and Setters //
}
#Entity
public class UserGroup {
#Id
#GeneratedValue(strategy = GenerationType.AUTO)
private long id;
private String name;
private String lookupKey;
#OneToMany(mappedBy="group",targetEntity=User.class, fetch=FetchType.EAGER)
#JsonManagedReference
private Collection users;
// Getters and Setters //
}
And also, here is the method in which it is implemented
public void summarizeGroupsByLookupKey(long ageFrom, long ageTo, List<String> lookupKeys, long numUsers){
CriteriaBuilder criteriaBuilder = entityManager.getCriteriaBuilder();
CriteriaQuery<Object[]> query = criteriaBuilder.createQuery(Object[].class);
Root<UserGroup> d = query.from(UserGroup.class);
Join<UserGroup, User> join = d.join("users");
Predicate pred1 = criteriaBuilder.between(join.get("age"), ageFrom, ageTo);
Expression<String> exp = d.get("lookupKey");
Predicate pred2 = exp.in(lookupKeys);
query.where(pred1, pred2);
query.multiselect(d.get("lookupKey"), criteriaBuilder.count(d.get("id")));
query.groupBy(d.get("lookupKey"));
query.having(criteriaBuilder.<Long>lessThan(criteriaBuilder.count(d.get("id")), numUsers));
List<Object[]> results = entityManager.createQuery(query).getResultList();
for(Object[] object : results){
System.out.println(object[0] + " " + object[1]);
}
}
By way of info...using Spring Boot 1.5.1 and all the default JPA, Hibernate, etc. from there.
Can a JPA expert offer some help? Thanks!

Answer 1

Change the multiselect part to use countDistinct(..)
query.multiselect(d.get("lookupKey")
,criteriaBuilder.countDistinct(d.get("id")));
and also having(..)
query.having(criteriaBuilder.<Long>lessThan(
criteriaBuilder.countDistinct(d.get("id")), numUsers)
);
Original query returned row per matching user in which rows userGroup.id was then multiplied.