We Keep Coding

jpa order records descending and select first record

I have a spring boot app connected to oracle DB.
I am trying to order a list of records and select the top most record.
I wrote a JPA query as below but it fails.
#Query("SELECT id FROM UploadedFile uploadedFile "
+ "WHERE uploadedFile.p_file_type = 'branch' "
+ "and p_file_status='Processed' "
+ "and p_is_file_process_with_error = 0 "
+ "order by c_created_date desc "
+ "FETCH FIRST 1 rows only ")
public String findLatestBranchCodeFile();
The error received was
creating bean with name 'uploadedFileRepo': Invocation of init method
failed; nested exception is java.lang.IllegalArgumentException:
Validation failed for query for method public abstract
java.lang.String
com.rhb.pintas.repo.UploadedFileRepo.findLatestBranchCodeFile()!
org.hibernate.hql.internal.ast.QuerySyntaxException: unexpected token:
FETCH near line 1, column 204 [SELECT id FROM
com.rhb.pintas.entities.UploadedFile uploadedFile WHERE
uploadedFile.p_file_type = 'branch' and p_file_status='Processed' and
p_is_file_process_with_error = 0 order by c_created_date desc FETCH
FIRST 1 rows only ] -> [Help 1]
The issue seems to be with fetch,not sure whats wrong.

It seems you are mixing HQL and native query dialects:
If this will be a naviveQuery (like most of the columns would mention), then replace the entity name with table name and add nativeQuery option. And because You are using only a single table, You can skip the alias name:
#Query("SELECT id FROM uploaded_file "
+ "WHERE p_file_type = 'branch' and p_file_status='Processed' and "
+ "p_is_file_process_with_error = 0 "
+ "order by c_created_date desc "
+ "FETCH FIRST 1 rows only ", nativeQuery = true)
public String findLatestBranchCodeFile();
If You want to keep it as a HQL, then replace all column names with entity property names, like p_file_type > fileType (I guess column names). Secondly You will need to pass a Pageable parameter, to replace Your 'Fetch first' statement.
You can find more materials here:
Bealdung
NativeQ
StackOverflow

You are trying to execute SQL query, in this case you need to add nativeQuery=true attribute to #Query annotation
UPD.
got confused because FETCH FIRST - is a SQL syntax, for JPA way please check another solution - it is possible to return list with at most one element.

I guess, you can try passing pagable to limit result set size and unlimit your query:
public String findLatestBranchCodeFile(Pageable pageable); // pass new PageRequest(0,1)

spring-data-jdbc: query containing entity with a 1-n relation

How can I write Queries for entities containing a 1-n reference?
Based on the spring-data-jdbc examples I will explain it with the following unit-test:
#Test
public void customQuery_ReferenceMultipleInstances() {
// prepare
LegoSet smallCar = createLegoSet("Small Car 01", 5, 12);
smallCar.setManual(new Manual("Just put all the pieces together in the right order", "Jens Schauder"));
smallCar.addModel("suv", "SUV with sliding doors.");
smallCar.addModel("roadster", "Slick red roadster.");
repository.save(smallCar);
// execute
List<LegoSet> actual = repository.findByName("Small Car 01");
Iterable<LegoSet> compare = repository.findAll();
// verify
assertThat(actual).as("same number of lego sets").hasSize(Lists.newArrayList(compare).size());
assertThat(actual.get(0).getModels()).as("same number of models").hasSize(Lists.newArrayList(compare).get(0).getModels().size());
assertThat(actual.get(0).getModels().get(0)).as("model must not be null").isNotNull();
assertThat(actual.get(0).getModels().get(0).getName()).as("model must have a name").isNotEmpty();
}
This models a LegoSet referencing 2 Models. The repository.findByName() is annotated with a custom query; the repository.findAll() is the standard spring-boot-data method of the CrudRepository (just as reference).
The custom query in version 1:
#Query("SELECT ls.id, ls.name, ls.min_age, ls.max_age, " +
"h.handbuch_id AS manual_handbuch_id, h.author AS manual_author, h.text AS manual_text " +
"FROM lego_set ls JOIN handbuch h ON ls.id = h.handbuch_id " +
"WHERE name = :name")
List<LegoSet> findByName(#Param("name") String name);
In this version the test fails w/
java.lang.AssertionError: [model must not be null]
Expecting actual not to be null
Okay, after that I add another JOIN to model:
#Query("SELECT ls.id, ls.name, ls.min_age, ls.max_age, " +
"h.handbuch_id AS manual_handbuch_id, h.author AS manual_author, h.text AS manual_text, " +
"m.* " +
"FROM lego_set ls JOIN handbuch h ON ls.id = h.handbuch_id " +
"JOIN model m ON ls.id = m.lego_set " +
"WHERE name = :name")
List<LegoSet> findByName(#Param("name") String name);
Now the test fails w/
java.lang.AssertionError: [same number of lego sets]
Expected size:<1> but was:<2> in:
<[LegoSet(id=1, name=Small Car 01, minimumAge=P5Y, maximumAge=P12Y, manual=Manual(id=1, author=Jens Schauder , text=Just put all the pieces together in the right order), models={suv=Model(name=suv, description=SUV with sliding doors.), roadster=Model(name=roadster, description=Slick red roadster.)}),
LegoSet(id=1, name=Small Car 01, minimumAge=P5Y, maximumAge=P12Y, manual=Manual(id=1, author=Jens Schauder
So how do I have to write that query correctly?

The first query is actually the correct one and it is working fine.
The problem is in your test. models is a Map, but you treat it like a List which compiles because a list index is also a valid map key.
If you change the last two assertions in your test like this they will succeed:
assertThat(actual.get(0).getModels().get("suv")).as("model must not be null").isNotNull();
assertThat(actual.get(0).getModels().get("suv").getName()).as("model must have a name").isNotEmpty();
// ----------------------------------------^
An alternative is to use the second query but with a custom ResultSetExtractor to collect multiple rows for the multiple models into a single LegoSet.

I'm in a similar situation, but i don't see what is different in my situation.
I have three models:
Client: id, name..
Project: id, name, clientId, List
ProjectMember
Here is my had coded query
#Query("""
select
project.id,
project.legacy_id,
project.code,
project.client_id,
project.is_archived,
project.name,
project.budget_type,
project.hours_budget_max,
project.money_budget_max
from project
join client on client.id = project.client_id
where client.name ilike '%:name%'
""")
List<Project> findByClientNameContainingIgnoreCase(final String name);
As you can see i'm looking for project where client name (using a join) match a pattern.
Executing the request by hand give me the expected results. But going through spring data jdbc give me no results

saveOrUpdate doesn't work in Spring Data with H2 database

I have for loop in my program where I save new objects to database. It looks like
for (String value: readvalue.readValue()) {
Value value= getValueForSomething(something);
System.out.println(value);
valueRepository.save(value);
}
And this fragment of code is executed every 30s and saving to database all values. Some values in database have two same fields and one other. How can I update values in h2 database instead of insert new?

I would suggest that within your for loop, create a method that checks to see if the object already exists in your H2 DB by querying for it using an unique identifier like an id. Use the following example RDB query as a reference:
private static final String PRODUCT_ALREADY_EXISTS_QUERY = "SELECT EXISTS(SELECT 1"
+ " FROM inventory.products "
+ " WHERE 1 = 1"
+ " AND id = :id)";
Then, if the record does exists, then call an update method that utilizes a query to UPDATE using the unique identifier. An RDB example query would be:
private static final String UPDATE_QUERY = "UPDATE inventory.products"
+ " SET (company_id, name, price, type, quantity, created_date, last_modified_date) = "
+ " (:companyId, :productName, :price, :productType, :quantity, :createdDate, :lastModifiedDateTime)"
+ " WHERE id = :id ";
If the record doesn't exist, then just create the record like you are.

Bind array param to native query

I have table product_spec_entry with following columns:
product_spec_id
commodity_spec_id
for one product_spec_id may be several commodity_spec_id, for example:
|product_spec_id | commodity_spec_id|
|----------------|------------------|
|1683 |1681 |
|1692 |1693 |
|1692 |1681 |
|1692 |1687 |
|1692 |1864 |
|1860 |1681 |
|1868 |1681 |
|1868 |1864 |
I want get all product_spec_id that have all commodity_spec_id are passed as parameter.
I wrote next query:
SELECT ps.product_spec_id, commodities
FROM (
SELECT
product_spec_id,
array_agg(commodity_spec_id) AS commodities
FROM system.product_spec_entry
GROUP BY product_spec_id) ps
WHERE Cast(ARRAY [1681, 1864] as BIGINT[]) <# Cast(ps.commodities as BIGINT[]);
It's work fine, and return expected result:
product_spec_id = 1692, 1868
I try use this query for JPA native query:
String query = "SELECT ps.product_spec_id " +
"FROM ( " +
" SELECT " +
" product_spec_id, " +
" array_agg(commodity_spec_id) AS commodities " +
" FROM system.product_spec_entry " +
" GROUP BY product_spec_id) ps " +
"WHERE CAST(ARRAY[:commoditySpecIds] AS BIGINT[]) <# CAST(ps.commodities AS BIGINT[])";
List<Long> commoditySpecsIds = commoditySpecs.stream().map(Spec::getId).collect(Collectors.toList());
List<BigInteger> productSpecIds = em.createNativeQuery(query).setParameter("commoditySpecIds", commoditySpecsIds)
.getResultList();
It does not work because I get array of record (ARRAY[(1692, 1868)]) instead array of bigint (ARRAY[1692, 1868])
How I should bind array param to my query? May be I can use more simple query for it.

Leave out the array[...] from your SQL:
WHERE CAST(:commoditySpecIds AS BIGINT[])
and then pass the list of IDs as a string that looks like this:
"{1,2,3,4}"
The default toString() for Lists usually returns something like: "[1,2,3]", so you could do something like this:
String literal = commoditySpecsIds.toString();
literal = "{" + literal.substring(1,literal.length() - 1) + "};
and then pass that to your obfuscation layer:
setParameter("commoditySpecIds", literal)

I'm exactly in the same situation. Hope #VladMihalcea can help us
Edit
I figure it out to do it with JPA. After reading the impementation of setParameter, i discovered something similar to UserType, the TypedParameterValue.
When you use
setParameter("commoditySpecIds", new TypedParameterValue(IntArrayType.INSTANCE, commoditySpecsIds))
Where IntArrayType.INSTANCE come from "hibernate-types" librairy provided by Vlad Mihalcea. Be carefull, the "commoditySpecsIds" must be an array, not a Collection.
Hope that helps

Other approcah for your case is unwrap the JPA provider session and use some methods form the JPA provider API.
If you are using hibernate, you can do the follow:
// unwrap hibenate session
final Session hibernateSession = em.unwrap(Session.class);
// create you SQL query in hibernate style
final SQLQuery sqlQuery = hibernateSession.createSQLQuery(sql);
And then set the parameter also using hibernate API
final Type customType = new CustomType(new ArrayTypeUser());
sqlQuery.setParameter("commoditySpecIds", value, customType);
Where "ArrayTypeUser" is a custom type that maps PostgresSQL array type to a Java array type.
This is not the best solution, but as you is are already using native queries, maybe for this particular case the best solution is skip the JPA standard API and use the JPA provide API.

Error: Cannot create TypedQuery for query with more than one return

I try to do the function searchBook with java and jpa. I have 2 classes which are Media and Book. Book extends Media. And I keep the data in the different table. I try to select the data from the query below:
TypedQuery<Media> query = em.createQuery(
"SELECT m.title, b.isbn, b.authors"
+ " FROM Book b, Media m" + " WHERE b.isbn = :isbn"
+ " OR lower(m.title) LIKE :title"
+ " OR b.authors LIKE :authors", Media.class);
query.setParameter("isbn", book.getisbn());
query.setParameter("title", "%" + book.getTitle().toLowerCase()
+ "%");
query.setParameter("authors", "%" + book.getAuthors() + "%");
bookList = query.getResultList();
But I got the error:
java.lang.IllegalArgumentException: Cannot create TypedQuery for query
with more than one return
This is the first time I use JPA. I can't find the the mistake.

As a workaround, to get entity composed by other entity attributes, you can create it within query, providing constructor for it.
Query :
TypedQuery<Media> query = em.createQuery("SELECT NEW package_name.Media(m.title, b.isbn, b.authors)"
+ " FROM Book b, Media m"
+ " WHERE b.isbn = :isbn"
+ " OR lower(m.title) LIKE :title"
+ " OR b.authors LIKE :authors", Media.class);
Entity :
public Media(String title, int isbn, String author){
//-- Setting appropriate values
}
I have provided sample, change the datatypes of the constructor accordingly.

Without goind into details about how Media and Book should be modeled, I will at least explain why you get this exception.
You're doing:
em.createQuery(someJPQL, Media.class);
This means: create a query using someJPQL, and this query will return instances of the Media entity.
But your JPQL is:
SELECT m.title, b.isbn, b.authors ...
So the query does not return entities of type Media. It returns three fields, from two different entities. There is no way your JPA engine could magically create instances of Media from these 3 columns. A query would return instances of Media if it looked like this:
select m from Media m ...

If you still want to use TypedQuery you can change the result type to Object[].
List<Object[]> results = entityManager
.createQuery("SELECT m.title, b.isbn, b.authors ...", Object[].class)
.getResultList();
Each Object[] in the List represents a row of data. It contains the selected values for that row in the order in which they were selected in the query. Element 0 is the title, element 1 is the ISBN, and element 2 is the authors. You'll likely need to cast those values if you want to use them in a meaningful way. Since the field values come from two different tables, you could store them in some kind of container object.
List<MediaContainer> mediaList = new ArrayList<>();
for (Object[] row : results) {
MediaContainer container = new MediaContainer();
container.setTitle((String) row[0]);
container.setIsbn((int) row[1]);
container.setAuthors((String) row[2]);
mediaList.add(container);
}

#WebUser instead of doing
List<EntityIDKey> companies =
getEntityManager().createQuery(sql, EntityIDKey.class).getResultList();
Try this :
List<EntityIDKey> companies =
(List<EntityIDKey>)getEntityManager().createQuery(sql).getResultList();
works for me.

if your are using Hibernate version < 4, you can meet this bug.
I go same problem with v3.5. Finally i had to use simple Query and cast each parameter manually
see other comments here : https://groups.google.com/forum/#!topic/axonframework/eUd1d4rotMY

I... remove
Media.class
of
createQuery
because you return more Entities in this source "SELECT m.title, b.isbn, b.authors"
Ex.:
TypedQuery<Media> query = em.createQuery(
"SELECT m.title, b.isbn, b.authors"
+ " FROM Book b, Media m" + " WHERE b.isbn = :isbn"
+ " OR lower(m.title) LIKE :title"
+ " OR b.authors LIKE :authors");

You can resolve the issue by doing this kind of query:
em.createQuery( "SELECT m"
+ " FROM Book b, Media m" + " WHERE b.isbn = :isbn"
+ " OR lower(m.title) LIKE :title"
+ " OR b.authors LIKE :authors", Media.class);
but this works only if you need fields only from one of the requested tables