What alternatives do I have to implement a union query using hibernate? I know hibernate does not support union queries at the moment, right now the only way I see to make a union is to use a view table.
The other option is to use plain jdbc, but this way I would loose all my example/criteria queries goodies, as well as the hibernate mapping validation that hibernate performs against the tables/columns.
You could use id in (select id from ...) or id in (select id from ...)
e.g. instead of non-working
from Person p where p.name="Joe"
union
from Person p join p.children c where c.name="Joe"
you could do
from Person p
where p.id in (select p1.id from Person p1 where p1.name="Joe")
or p.id in (select p2.id from Person p2 join p2.children c where c.name="Joe");
At least using MySQL, you will run into performance problems with it later, though. It's sometimes easier to do a poor man's join on two queries instead:
// use set for uniqueness
Set<Person> people = new HashSet<Person>((List<Person>) query1.list());
people.addAll((List<Person>) query2.list());
return new ArrayList<Person>(people);
It's often better to do two simple queries than one complex one.
EDIT:
to give an example, here is the EXPLAIN output of the resulting MySQL query from the subselect solution:
mysql> explain
select p.* from PERSON p
where p.id in (select p1.id from PERSON p1 where p1.name = "Joe")
or p.id in (select p2.id from PERSON p2
join CHILDREN c on p2.id = c.parent where c.name="Joe") \G
*************************** 1. row ***************************
id: 1
select_type: PRIMARY
table: a
type: ALL
possible_keys: NULL
key: NULL
key_len: NULL
ref: NULL
rows: 247554
Extra: Using where
*************************** 2. row ***************************
id: 3
select_type: DEPENDENT SUBQUERY
table: NULL
type: NULL
possible_keys: NULL
key: NULL
key_len: NULL
ref: NULL
rows: NULL
Extra: Impossible WHERE noticed after reading const tables
*************************** 3. row ***************************
id: 2
select_type: DEPENDENT SUBQUERY
table: a1
type: unique_subquery
possible_keys: PRIMARY,name,sortname
key: PRIMARY
key_len: 4
ref: func
rows: 1
Extra: Using where
3 rows in set (0.00 sec)
Most importantly, 1. row doesn't use any index and considers 200k+ rows. Bad! Execution of this query took 0.7s wheres both subqueries are in the milliseconds.
Use VIEW. The same classes can be mapped to different tables/views using entity name, so you won't even have much of a duplication. Being there, done that, works OK.
Plain JDBC has another hidden problem: it's unaware of Hibernate session cache, so if something got cached till the end of the transaction and not flushed from Hibernate session, JDBC query won't find it. Could be very puzzling sometimes.
I have to agree with Vladimir. I too looked into using UNION in HQL and couldn't find a way around it. The odd thing was that I could find (in the Hibernate FAQ) that UNION is unsupported, bug reports pertaining to UNION marked 'fixed', newsgroups of people saying that the statements would be truncated at UNION, and other newsgroups of people reporting it works fine...
After a day of mucking with it, I ended up porting my HQL back to plain SQL, but doing it in a View in the database would be a good option. In my case, parts of the query were dynamically generated, so I had to build the SQL in the code instead.
I have a solution for one critical scenario (for which I struggled a lot )with union in HQL .
e.g. Instead of not working :-
select i , j from A a , (select i , j from B union select i , j from C) d where a.i = d.i
OR
select i , j from A a JOIN (select i , j from B union select i , j from C) d on a.i = d.i
YOU could do in Hibernate HQL ->
Query q1 =session.createQuery(select i , j from A a JOIN B b on a.i = b.i)
List l1 = q1.list();
Query q2 = session.createQuery(select i , j from A a JOIN C b on a.i = b.i)
List l2 = q2.list();
then u can add both list ->
l1.addAll(l2);
A view is a better approach but since hql typically returns a List or Set... you can do list_1.addAll(list_2). Totally sucks compared to a union but should work.
Perhaps I had a more straight-forward problem to solve. My 'for instance' was in JPA with Hibernate as the JPA provider.
I split the three selects (two in a second case) into multiple select and combined the collections returned myself, effectively replacing a 'union all'.
Hibernate 6 added support for UNION.
So, you can now use UNION in JPQL queries like this:
List<String> topics = entityManager.createQuery("""
select c.name as name
from Category c
union
select t.name as name
from Tag t
""", String.class)
.getResultList();
And you can also also use UNION ALL if there are no duplicates to be removed:
List<String> topics = entityManager.createQuery("""
select c.name as name
from Category c
union all
select t.name as name
from Tag t
""", String.class)
.getResultList();
Besides UNION, you can also use EXCEPT and INTERSECT.
I too have been through this pain - if the query is dynamically generated (e.g. Hibernate Criteria) then I couldn't find a practical way to do it.
The good news for me was that I was only investigating union to solve a performance problem when using an 'or' in an Oracle database.
The solution Patrick posted (combining the results programmatically using a set) while ugly (especially since I wanted to do results paging as well) was adequate for me.
Here is a special case, but might inspire you to create your own work around. The goal here is to count the total number of records from two different tables where records meet a particular criteria. I believe this technique will work for any case where you need to aggregate data from across multiple tables/sources.
I have some special intermediate classes setup, so the code which calls the named query is short and sweet, but you can use whatever method you normally use in conjunction with named queries to execute your query.
QueryParms parms=new QueryParms();
parms.put("PROCDATE",PROCDATE);
Long pixelAll = ((SourceCount)Fetch.row("PIXEL_ALL",parms,logger)).getCOUNT();
As you can see here, the named query begins to look an aweful lot like a union statement:
#Entity
#NamedQueries({
#NamedQuery(
name ="PIXEL_ALL",
query = "" +
" SELECT new SourceCount(" +
" (select count(a) from PIXEL_LOG_CURR1 a " +
" where to_char(a.TIMESTAMP, 'YYYYMMDD') = :PROCDATE " +
" )," +
" (select count(b) from PIXEL_LOG_CURR2 b" +
" where to_char(b.TIMESTAMP, 'YYYYMMDD') = :PROCDATE " +
" )" +
") from Dual1" +
""
)
})
public class SourceCount {
#Id
private Long COUNT;
public SourceCount(Long COUNT1, Long COUNT2) {
this.COUNT = COUNT1+COUNT2;
}
public Long getCOUNT() {
return COUNT;
}
public void setCOUNT(Long COUNT) {
this.COUNT = COUNT;
}
}
Part of the magic here is to create a dummy table and insert one record into it. In my case, I named it dual1 because my database is Oracle, but I don't think it matters what you call the dummy table.
#Entity
#Table(name="DUAL1")
public class Dual1 {
#Id
Long ID;
}
Don't forget to insert your dummy record:
SQL> insert into dual1 values (1);
As Patrick said, appending the LISTs from each SELECT would be a good idea but remember that it acts like UNION ALL. To avoid this side effect, just control if the object is already added in final collection or not. If no, then add it.
Something else that you should care about is that if you have any JOIN in each SELECT, the result would be a list of object array(List<Object[]>) so you have to iterate over it to only keep the object that you need.
Hope it works.
Related
I want to get the number of results of a query in Spring Data Jpa, using a non-native #Query method. It consists of a basic group by plus a having clause.
My plain query looks like this (analogous example):
select count(*) from (
select 1 from table t
where t.field_a = 1
group by t.id
having count(*) = 2) a;
Since Hibernate 5 does not allow subqueries in the form clause, I have to find a workaround for that. The only one I found is very inefficient as per the query plan:
select count(*) from table t
where t.field_a = 1 and
2 = (select count(*) from table temp where temp.id = t.id);
Is there a way to write a Spring Data JPA query that's as efficient as the first one? I can think of no solution rather than selecting the inner query and taking its size() in java, but that can produce issues due to a ton of redundant data passing through the network.
There is no easy solution to count the results of a subquery in JPA but the a workaround is proposed here https://arjan-tijms.omnifaces.org/2012/06/counting-rows-returned-from-jpa-query.html.
The principle is to build a native query based on the initial Jpa subselect query.
This does the job if you accept to count the elements in java !
Query q = em.createQuery(
"select 1 from table t where field_a = 1 " +
"group by t.id having count(*) = 2");
int count = q.getResultList().size();
(performances depending on the number of lines returned, but the projection is very light : 1)
I have a java method which finds all children (leaves) of a parent. It works ok.but there is performance issue. How can i make it work faster? When the children gets large it becomes problematic. Here is the method:
Private void getChildren(String orgId, List<String> children, List<String> organs, int j){
Query query= entityManager.createNativeQuery("select c.organization_id from organization c where c.parent_org_id="+orgId);
if(query.getResultList().size()>0)
organs.addAll((List<String>) query.getResultList());
else
children.add(orgId);
for(int i=j+1; i<organs.size();i++){
j=i;
query=entityManager.createNativeQuery("select c.organization_id from organization c where c.parent_org_id="+String.valueOf(organs.get(i)));
if(query.getResultList().size()>0)
organs.addAll((List<String>) query.getResultList());
else
children.add(String.valueOf(organs.get(i)));
}
}
I call this method like
getChildren("68",new ArrayList<>(),new ArrayList<>(),-1)
I have oracle as database so if it is better to acheive this via oracle query please note me.
By the way my spring version doesn't support streams!
You basically have a 1+N select problem (issue a single query for each id in your collection). Don't do this, instead use the IN clause and issue a single query!.
Another thing you shouldn't be doing is use concat for your query creation, never do this! What you should do is just return the result and use a query with an IN clause.
String query = "select c.organization_id from organization c where c.parent_org_id IN (:ids))
return em.createNativeQuery(query).setParameters("ids", organs);
or even better use a single query altogether. You can use a self join for that)
String query = "SELECT c1.organization_id " +
"FROM organization c1 LEFT OUTER JOIN organization c2 " +
"ON c1.parent_org_id=c2.organization_id " +
"WHERE c2.parent_org_id=?";
return em.createNativeQuery(query, String.class).setParameter(1, orgId).getResultList();
Judging from your solution I would strongly recommend studying SQL and JPA.
Try it in a big query
select c.organization_id from organization, (
select c.organization_id as sub1 from organization c where c.parent_org_id="+orgId +") as x_ where organization.parent_org_id = x_.sub1
The reason why it is slow, is that you query the database for every child. This is very inefficient, I assume even the load on the database will be smaller.
Answer provided by Deinum only gets the direct children. Whereas my java method returns all children including nested ones.
I have found a query for that which is much more optimized than my java method. It uses temporary tables:
with temp_tbl (id,parent_id) as
(
select organization_id,parent_org_id from organization where parent_org_id =?
union all
select c.organization_id ,c.parent_org_id from organization c join temp_tbl t on t.id=c.parent_org_id and t.parent_id !=t.id
)
Select to_char(g.id) from temp_tbl g
At the end i return to_char(g.id) to make it a string instead of big decimal. Because em.createNativeQuery(query, String.class) didn't turn it to string and threw error for me.
So just use em.createNativeQuery(query)
I want to search in 16 different tables, but I don't wanna repeat the "select from DB" 16 times; I think that's not really help in performance!!!
I am using:
query="SELECT * FROM table1, table2,..., table16 WHERE id=?";
Is it correct ??
my problem is how to separate between data of tables ??
also maybe I can get from one table two or more results for one "id"; So I want to know which data is from which table !!
.
Best regards,
Your query will not work, because you are trying to join those multiple tables, whereas what you want to do is search (filter) those 16 tables.
You could use a union all to do this in a single query:
select xxx, 'table1' as source_table
from table1
where id = ?
union all
select xxx, 'table2' as source_table
from table2
where id = ?
and so on. The second derived field source_table can be used to determine which table returned which result.
You have to list all fields using aliases for fields with same name, and prefix with table names.
For example :
query = "SELECT table1.id as id_1, table2.id as id_2, ... WHERE id_1 = 23"
Probably a very long query to write, but you have solution to generate and paste it : You can do this for example with FlySpeed SqlQuery (free for personal use)
FlySpeed SqlQuery will generate all aliases for you, and automatically prefix with table names.
A little clarification would help. If all 16 tables have the same fields and you want them in a continuous list, you can use UNION as suggested above. On the other hand, if there are only a few fields that match and you want to compare the values for each table side-by-side, you'll want to use joins and provide aliases with the table names, as also suggested above.
However, looking at the snippet of code you've provided, I'm going to guess that you're either building some kind of stored procedure or else implementing SQL in some other language. If that's the case, how about loading your table names into an array and using a for loop to build the query, such as the following psuedo-code:
tableList = ["table1", "table2"...]
fieldnames = ["field1", "field2"...]
query = "SELECT "
for i = 0 to count(tableList):
for j = 0 to count(fieldnames):
query = query + tablelist[i] + "." + fieldnames[j] + ", "
j++
i++
query = query + "FROM "
for i = 0 to count(tableList):
query = query + tableList[i] + ", "
i++
query = query + "WHERE " ...
And so forth. Much of this depends on what exactly you're looking to do, how often you're looking to do it, and how often the variables (like which tables or fields you're using) are going to change.
I'm trying to write a HQL/Criteria/Native SQL query that will return all Employees that are assigned to a list of Projects. They must be assigned to all Projects in order to be selected.
An acceptable way of achieving this with native SQL can be found in the answer to this question: T-SQL - How to write query to get records that match ALL records in a many to many join:
SELECT e.id
FROM employee e
INNER JOIN proj_assignment a
ON e.id = a.emp_id and a.proj_id IN ([list of project ids])
GROUP BY e.id
HAVING COUNT(*) = [size of list of project ids]
However, I want to select all fields of Employee (e.*). It's not possible to define SQL grouping by all the columns(GROUP BY e.*), DISTINCT should be used instead. Is there a way to use DISTINCT altogether with COUNT(*) to achieve what I want?
I've also tried using HQL to perform this query. The Employee and ProjectAssignment classes don't have an association, so it's not possible to use Criteria to join them. I use a cross join because it's the way to perform a Join without association in HQL. So, my HQL looks like
select emp from Employee emp, ProjectAssignment pa
where emp.id = pa.empId and pa.paId IN :list
group by emp having count(*) = :listSize
However, due to a bug in Hibernate, GROUP BY entity does not work. The SQL it outputs is something like group by (emptable.id).
Subquerying the assignment table for each project (dynamically adding and exists (select 1 from proj_assignment pa where pa.emp_id=e.id and pa.proj_id = [anId]) for each project in the list) is not an acceptable option.
Is there a way to write this query properly, preferrably in HQL (in the end I want a List<Employee>), without modifying mappings and without explicitly selecting all columns in the native SQL ?
EDIT: I'm using Oracle 10g and hibernate-annotations-3.3.1.GA
How about:
select * from employee x where x.id in(
SELECT e.id
FROM employee e
INNER JOIN proj_assignment a
ON e.id = a.emp_id and a.proj_id IN ([list of project ids])
GROUP BY e.id
HAVING COUNT(*) = [size of list of project ids]
)
I've found an alternative way to achieve this in HQL, it's far more inefficient than what I'd like, (and than what is really possible without that nasty bug) but at least it works. It's better than repeating subselects for each project like
and exists (select 1 from project_assignment pa where pa.id = someId and pa.emp_id = e.id)
It consists of performing a self-join subquery in order to find out, for each of the Employees, how many of the projects in the list they are assigned to, and restrict results to only those that are in all of them.
select e
from Employee
where :listSize =
(select distinct count(*)
from Employee e2, ProjectAssignment pa
where
e2.id = pa.id_emp and
e.id = e2.id
and pa.proj_id IN :projectIdList
)
I'm new to ORM stuff and I need some help understanding something.
Let's assume I have the following standard SQL query:
SELECT *, COUNT(test.testId) AS noTests FROM inspection
LEFT JOIN test ON inspection.inspId = test.inspId
GROUP BY inspection.inspId
which I want to use in JPA.
I have an Inspection entity with a one-to-many relationship to a Test entity. (an inspection has many tests)
I tried writing this in JPQL:
Query query = em.createQuery("SELECT insp, COUNT(???what???) " +
"FROM Inspection insp LEFT JOIN insp.testList " +
"GROUP BY insp.inspId");
1) How do I write the COUNT clause? I'd have to apply count to elements from the test table but testList is a collection, so I can't do smth like COUNT(insp.testList.testId)
2) Assuming 1 is resolved, what type of object will be returned. It will definitely not be an Inspection object... How do I use the result?
You can give an alias to the joined entity (with AS)
You can create either a new object, or a List with the returned values
So:
SELECT new com.yourproject.ResultHolder(insp, COUNT(test.testId))
FROM Inspection insp LEFT JOIN insp.testList AS test GROUP BY insp.inspId
Or
SELECT new list(insp, COUNT(test.testId))
FROM Inspection insp LEFT JOIN insp.testList AS test GROUP BY insp.inspId
The result is then accessible either as an instance of ResultHolder, or as a java.util.List, where the insp is list.get(0), and the count is list.get(1)