How to write JPQL query using multiple members of same collection? - jpa

I have a noticication entity that has OneToMany relationship with its parameters, which is a list of NotificationParamEntity objects.
The code for both classes looks like:
// Notification Entity
#Entity
#Table (name = "NOTIFICATIONS")
public class NotificationEntity {
......
#OneToMany (mappedBy = "notification")
private List<NotificationParamEntity> params;
......
}
// Notification Parameter Entity
#Entity
#Table (name = "NOTIFICATIONS_PARAM")
public class NotificationParamEntity {
......
#Column (name = "KEY", length = 40, nullable = false)
#Enumerated (EnumType.STRING)
private NotificationParameterEnum key;
#Column (name = "VALUE", length = 4000, nullable = false)
private String value;
#ManyToOne
#JoinColumn (name = "NOTIFICATION_ID", nullable = false)
private NotificationEntity notification;
......
}
Now I can use the query below to get the notification that has a parameter named "P1" and with a value "V1":
SELECT DISTINCT anEntity FROM NotificationEntity anEntity, IN
(anEntity.params) p WHERE p.key = "P1" AND p.value = 'V1'
But when I want to find out the notification that has two specified parameters(P1=V1 and P2=V2), my query below failed to find anything:
SELECT DISTINCT anEntity FROM NotificationEntity anEntity, IN
(anEntity.params) p WHERE p.key = "P1" AND p.value = 'V1' AND p.key = "P2" AND p.value = "V2"
I can understand why it doesn't work: there is no parameter that can have two different keys, so the query return nothing.
But how to make this work? Assume I have a notification entity that has two parameters, one is named P1 and value is V1, the other one is P2 and the value is V2, how can I find this notification entity with JPQL query?

Try something like this:
SELECT n FROM NotificationEntity n WHERE EXISTS
(SELECT p FROM NotificationParamEntity p WHERE p.key = 'P1' AND p.value = 'V1'
AND p.notificationEntity = n)
AND EXISTS
(SELECT p2 FROM NotificationParamEntity p2 WHERE p2.key = 'P2' AND p2.value = 'V2'
AND p2.notificationEntity = n)
Note that it requires a reference from NotificationParamEntity to NotificationEntity (I don't see that column in the snippet of your code, but you should have it, a #ManyToOne).

Related

EF Core - how to select count from child table based on foreign key

TableB has a field TableAId which is linked to the Id of TableA. I want to select the count from TableB based on TableAId like -
SELECT *,
(SELECT COUNT(*) FROM tableB where tableB.TableAId = tableA.Id) as count
FROM tableA
So far I have the code:
var data = _context.TableA.AsQueryable();
...
data = data.Select(l => l.TableAId= p.Id).Count();
But on the last line, p is not recognized as a variable.
How do I make this work?
EDIT :
my original query is quiet complex and already filtering data
var data = _context.TableA.AsQueryable();
data = data.Include(p => p.SomeClassA)
.Include(p => p.SomeClassB);
data = data.Where(p => p.Id == somevalue);
data = data.Where(p => p.SomeClassA.Name.Contains(someothervalue));
data = data.Where(p => p.SomeClassA.SomeField.Contains(yetanothervalue));
I tried adding this but it cannot compile
(TableAId & Count do not exist):
data = data.Join(
_context.TableB,
groupByQuery => groupByQuery.TableAId ,
TableA => TableA.Id,
(groupByQuery, TableAItem) => new
{
TableAId = groupByQuery.Id,
Count = groupByQuery.Count,
TableAItem = TableAItem
}
);
If you are just interested in count, then do the following:
var data = _context.TableB.AsQueryable();
var groupByCountQuery = data.GroupBy(a=>a.TableAId, (tableAId, tableBItems) => new
{
TableAId = tableAId,
Count = tableBItems.Count()
});
var result = groupByCountQuery.ToList(); // or change to ToListAsync()
This will give you the count based on TableAId.
If you need the tableA items as well in the result, following ca be done:
var groupByCountQuery = data.GroupBy(a=>a.TableAId, (tableAId, tableBItems) => new
{
TableAId = tableAId,
Count = tableBItems.Count()
}).Join(_context.TableA,
groupByQuery => groupByQuery.TableAId,
tableA => tableA.Id,
(groupByQuery , tableA) => new {
TableAId = groupByQuery.TableAId,
Count = groupByQuery.Count,
TableAItem = tableA
} );
Let's say TableA has properties - Id, FieldA1 and FieldA2.
First, you have to include TableB in your query so that you can take its count, like -
var data = _context.TableA.Include(p=> p.TableB).AsQueryable();
Then in the Select method you have to create a new object with TableA's properties and TableB's count, like -
var list = data.Select(p =>
new
{
Id = p.Id,
A1 = p.FieldA1,
A2 = p.FieldA2,
Count = p.OrderLines.Count
}).ToList();
Notice, this is assigned to a new variable list. That is because it does not return a list of TableA, it returns a list of an anonymous object with properties - Id, A1, A2 and Count. Therefore, you cannot assign it to the previously declared data variable, because data is of type IQueryable<TableA>.
Alternatively, you can declare a class to hold the data values, like -
public class MyData
{
public int Id { get; set; }
public string A1 { get; set; }
public string A2 { get; set; }
public int Count { get; set; }
}
and use it like -
var list = data.Select(p =>
new MyData
{
Id = p.Id,
A1 = p.FieldA1,
A2 = p.FieldA2,
Count = p.OrderLines.Count
}).ToList();
If you want to start from TableA, you can use the following linq query:
var data = _context.TableAs.AsQueryable();
var x = (from a in data
join b in _context.TableBs on a.Id equals b.TableAId
group a.TableB by a.Id into g
select new
{
TableAId = g.Key,
TableBItem = g.FirstOrDefault(),
Count = g.Count()
}).ToList();
Result:

Inserting parent/child rows in Hibernate with autoincrement index throws error

I have the following Hibernate classes in Scala, where one Group has many Items. Note that the #Id of the Group class has an autoincrement annotation.
#Entity
#Table(name = "items")
class Item extends Serializable {
#Id
#ManyToOne
#JoinColumn(name="group_sk", nullable=false)
var group: Group = _
#Id
var index: Int = _
var name: String = _
def canEqual(a: Any) = a.isInstanceOf[Item]
override def equals(that: Any): Boolean =
that match {
case that: Item => that.canEqual(this) && this.hashCode == that.hashCode
case _ => false
}
override def hashCode: Int = {
val prime = 31
var result = 1
result = prime * result + group.sk;
result = prime * result + index
return result
}
}
#Entity
#Table(name = "groups")
class Group {
#Id
#GeneratedValue(strategy = GenerationType.SEQUENCE, generator = "group_generator")
#SequenceGenerator(name="group_generator",
sequenceName = "GroupSeq", allocationSize = 1)
var sk: Int = _
#Column(name = "group_name")
var name: String = _
#OneToMany
#JoinColumn(name="group_sk")
var items: java.util.List[Item] = _
}
I try to insert one group with a related item, where both the Group should have an auto incremented Id:
session.beginTransaction
val group = new Group
group.name = "Group name"
group.items = new java.util.ArrayList[Item]()
val item1 = new Item
item1.group = group
item1.index = 1
item1.name = "Item 1"
group.items.add(item1)
session.save(group)
session.getTransaction.commit
The exception I get is
Caused by: javax.persistence.OptimisticLockException: Batch update
returned unexpected row count from update [0]; actual row count: 0;
expected: 1
And the Hibernate sql log shows:
Hibernate: select GroupSeq.nextval from dummy
Hibernate: insert into groups (group_name, sk) values (?, ?)
Hibernate: update items set group_sk=? where index=? and group_sk=?
Note that the last update statement doesn't make sense because you won't update the value of group_sk where the column is also in the condition. Moreover, there's no insert statement of the items table.
How to fix this problem?
You need to set cascade type on your OneToMany relationship to tell Hibernate how to mange Items collection of the Group class. So for example something like this should work:
#OneToMany(cascade=CascadeType.ALL)
#JoinColumn(name="group_sk")
var items: java.util.List[Item] = _

Criteria API - multiselect - Tuple which contains other Tuple/Collection

I have following entities:
class Person {
int id;
String name;
List<Address> addresses
}
class Address {
int id;
String city;
}
I try to prepare query (based on Criteria-API) where result (each tuple in result List) will contain three elements:
persron.id
person.name
person.addresses <-- collection with 0 or more elements
(1, "Peter", Collection{2,3}) or
(1, "Peter", Tuple{2,3})
I have tried something like this:
EntityManager em;
CriteriaBuilder qb = em.getCriteriaBuilder();
CriteriaQuery<Tuple> cq = qb.createTupleQuery()
Root<Person> root = cq.from(Person.class);
ListJoin<PersonalData, Address> join = (ListJoin)root.join("addresses", JoinType.LEFT);
cq.multiselect(root.get("id"), root.get("name"), join.get("id"));
TypedQuery<Tuple> tq = em.createQuery(cq);
List<Tuple> result = tq.getResultList();
But received result is different that expected :(
For model: Person(1, "Peter")
which has two addresses
Address(2, "London");
Address(3, "Paris");
my result list is something like Cartesian product:
(1, "Peter", 2), (1, "Peter", 3)
Is is possible to receive result which was requested at the beginning of this post?
No, it can not work the way you are expecting, as the join between person and address will result in a single address per row, and JPA will return the data in a similar format. One person.id, person.name, person.address in each tuple. Why not just return the Person instance and use that?

An anonymous type cannot have multiple properties with the same name

I want to bind gridview through entity framework but it throws error
like-
An anonymous type cannot have multiple properties with the same name Entity Framwrok
Here is my method.
public void UserList(GridView grdUserList)
{
using (TreDbEntities context = new TreDbEntities())
{
var query =( from m in context.aspnet_Membership
from u in context.aspnet_Users
join usr in context.Users
on new { m.UserId, u.UserId }
equals new { usr.MembershipUserID, usr.UserId }
into UserDetails
from usr in UserDetails
select new {
CreationDate = m.CreateDate,
email = m.Email,
UserName = u.LoweredUserName,
Name = usr.FirstName + usr.LastNameLastName,
Active=usr.IsActive
}).ToList();
}
}
It shows error here. usr.UserId.
The direct issue is in the anonymous type new { m.UserId, u.UserId }: the same name twice. You can fix that by giving explicit property names, for example: new { u1 = m.UserId, u2 = u.UserId }.
But then the next issue will be that both anonymous types that define the join will not have the same property names, so the final fix is this:
public void UserList(GridView grdUserList)
{
using (TreDbEntities context = new TreDbEntities())
{
var query =( from m in context.aspnet_Membership
from u in context.aspnet_Users
join usr in context.Users
on new { u1 = m.UserId, u2 = u.UserId }
equals new { u1 = usr.MembershipUserID, u2 = usr.UserId }
into UserDetails
from usr in UserDetails
select new { CreationDate = m.CreateDate,
email = m.Email,
UserName = u.LoweredUserName,
Name = usr.FirstName + " " + usr.LastName,
Active = usr.IsActive
}
).ToList();
}
}
#Gert answer is correct. Just want to show simpler solution - give name only to first UserId property:
on new { MembershipUserID = m.UserId, u.UserId }

Convert datetime to a formatted string inside a LINQ-to-entities query

How can I convert DateTime into a formatted string?
This is the line in the following query that needs help:
StartDate = string.Format("{0:dd.MM.yy}", p.StartDate)
The whole query:
var offer = (from p in dc.CustomerOffer
join q in dc.OffersInBranch
on p.ID equals q.OfferID
where q.BranchID == singleLoc.LocationID
let value = (p.OriginalPrice - p.NewPrice) * 100 / p.OriginalPrice
orderby value descending
select new Offer()
{
Title = p.OfferTitle,
Description = p.Description,
BestOffer = value,
ID = p.ID,
LocationID = q.BranchID,
LocationName = q.CustomerBranch.BranchName,
OriginalPrice = SqlFunctions.StringConvert((decimal)p.OriginalPrice),
NewPrice = SqlFunctions.StringConvert((decimal)p.NewPrice),
StartDate = string.Format("{0:dd.MM.yy}", p.StartDate)
}).First();
I get the following error message:
LINQ to Entities does not recognize the method 'System.String ToString(System.String)' method, and this method cannot be translated into a store expression.
Another option is using SqlFunctions.DateName, your code will be like this:
var offer = (from p in dc.CustomerOffer
join q in dc.OffersInBranch
on p.ID equals q.OfferID
where q.BranchID == singleLoc.LocationID
let value = (p.OriginalPrice - p.NewPrice) * 100 / p.OriginalPrice
orderby value descending
select new
{
Title = p.OfferTitle,
Description = p.Description,
BestOffer = value,
ID = p.ID,
LocationID = q.BranchID,
LocationName = q.CustomerBranch.BranchName,
OriginalPrice = SqlFunctions.StringConvert((decimal)p.OriginalPrice),
NewPrice = SqlFunctions.StringConvert((decimal)p.NewPrice),
StartDate = SqlFunctions.DateName("day", p.StartDate) + "/" + SqlFunctions.DateName("month", p.StartDate) + "/" + SqlFunctions.DateName("year", p.StartDate)
})
I found it useful if you don't want to add an extra select new block.
EDIT: Now that I understand the question, I'm giving it another shot :)
var offer = (from p in dc.CustomerOffer
join q in dc.OffersInBranch
on p.ID equals q.OfferID
where q.BranchID == singleLoc.LocationID
let value = (p.OriginalPrice - p.NewPrice) * 100 / p.OriginalPrice
orderby value descending
select new
{
Title = p.OfferTitle,
Description = p.Description,
BestOffer=value,
ID=p.ID,
LocationID=q.BranchID,
LocationName=q.CustomerBranch.BranchName,
OriginalPrice=SqlFunctions.StringConvert((decimal)p.OriginalPrice),
NewPrice=SqlFunctions.StringConvert((decimal)p.NewPrice),
StartDate=p.StartDate
})
.ToList()
.Select(x => new Offer()
{
Title = x.OfferTitle,
Description = x.Description,
BestOffer=value,
ID=x.ID,
LocationID=x.BranchID,
LocationName=x.CustomerBranch.BranchName,
OriginalPrice=x.OriginalPrice,
NewPrice=x.NewPrice,
StartDate=x.StartDate.ToString("dd.MM.yy")
}).First();
I know it's a bit long, but that's the problem with Linq To SQL.
When you use linq, the database call isn't executed until you use something such as ToList() or First() that results in actual objects. Once that SQL call is executed by the first .First() call, you're now working with .NET types, and can use DateTime stuff.
I ended up using the sql function FORMAT; here's a simplified version of this implementation:
https://weblogs.asp.net/ricardoperes/registering-sql-server-built-in-functions-to-entity-framework-code-first
First you need to define the function in EF:
public class FormatFunctionConvention : IStoreModelConvention<EdmModel>
{
public void Apply(EdmModel item, DbModel model)
{
var payload = new EdmFunctionPayload
{
StoreFunctionName = "FORMAT",
Parameters = new[] {
FunctionParameter.Create("value", PrimitiveType.GetEdmPrimitiveType(PrimitiveTypeKind.DateTime), ParameterMode.In),
FunctionParameter.Create("format", PrimitiveType.GetEdmPrimitiveType(PrimitiveTypeKind.String), ParameterMode.In)
},
ReturnParameters = new[] {
FunctionParameter.Create("result", PrimitiveType.GetEdmPrimitiveType(PrimitiveTypeKind.String), ParameterMode.ReturnValue)
},
Schema = "dbo",
IsBuiltIn = true
};
item.AddItem(EdmFunction.Create("FORMAT", "CodeFirstDatabaseSchema", item.DataSpace, payload, null));
}
}
Then define it as C# methods:
public static class SqlFunctions
{
[DbFunction("CodeFirstDatabaseSchema", "FORMAT")]
public static String Format(this DateTime value, string format)
{
return value.ToString(format);
}
[DbFunction("CodeFirstDatabaseSchema", "FORMAT")]
public static String Format(this DateTime? value, string format)
{
return value?.ToString(format);
}
}
Register it in your DbContext:
public class SqlDb : DbContext
{
protected override void OnModelCreating(DbModelBuilder modelBuilder)
{
base.OnModelCreating(modelBuilder);
modelBuilder.Conventions.Add(new FormatFunctionConvention());
}
}
And finally, you can call it like so:
var x = db.MyItems.Select(i => new { FormattedDate = SqlFunctions.Format(i.MyDate, "MM/dd/yyyy") }).ToArray();
That is what we did, we added a new function to the class and we query the date as normal in the query:
[ComplexType]
public class Offer
{
public DateTime StartDate
{
get;
set;
}
public String Title
{
get;
set;
}
/*Other fields*/
.
.
.
public string FormattedDate(string format)
{
return Date.ToString(format);
}
}
var offer = (from p in dc.CustomerOffer
join q in dc.OffersInBranch
on p.ID equals q.OfferID
where q.BranchID == singleLoc.LocationID
let value = (p.OriginalPrice - p.NewPrice) * 100 / p.OriginalPrice
orderby value descending
select new Offer()
{
Title = p.OfferTitle,
Description = p.Description,
BestOffer = value,
ID = p.ID,
LocationID = q.BranchID,
LocationName = q.CustomerBranch.BranchName,
OriginalPrice = SqlFunctions.StringConvert((decimal)p.OriginalPrice),
NewPrice = SqlFunctions.StringConvert((decimal)p.NewPrice),
StartDate = p.StartDate
}).First();
Then you can just call the FormattedDate field passing the desired format.
edit1.Text = offer.FormattedDate("dd.MM.yy");
Or can can define it as a field with just the getter:
public string FormattedDate
{
get { return Date.ToString("dd.MM.yy") };
}
edit1.Text = offer.FormattedDate;
In case you class is an Entity, you need to declare a new partial of that class and add the field.
Hope this help someone.
In vb (valid to c# too changing syntax):
Imports System.Data.Entity
...
query.Select(Function(x) New MyObject With {
...
.DateString = DbFunctions.Right("00" & x.DateField.Day, 2) & "/" & DbFunctions.Right("00" & x.DateField.Month, 2) & "/" & x.DateField.Year
...
}).ToList()
Note: ToList(), ToEnumerable() are not the way because its executes a query, the user wants linq to sql..
The easiest and most efficient way I have found to do string formats on numeric or datetime objects is by using string interpolation. It will bring back the actual DateTime/int/float/double/etc.. objects in the SQL query, and then client side it will do the string format during projection. I modified your query below, note how OriginalPrice, NewPrice, and StartDate are converted:
var offer = (from p in dc.CustomerOffer
join q in dc.OffersInBranch
on p.ID equals q.OfferID
where q.BranchID == singleLoc.LocationID
let value = (p.OriginalPrice - p.NewPrice) * 100 / p.OriginalPrice
orderby value descending
select new Offer()
{
Title = p.OfferTitle,
Description = p.Description,
BestOffer = value,
ID = p.ID,
LocationID = q.BranchID,
LocationName = q.CustomerBranch.BranchName,
OriginalPrice = $"{p.OriginalPrice:C2}",
NewPrice = $"{p.NewPrice:C2}",
StartDate = $"{p.StartDate:dd.MM.yy}"
}).First();
if it's a datetime you need to use the .ToShortDateString(). But you also need to declare it AsEnumerable().
var offer = (from p in dc.CustomerOffer.AsEnumerable()
join q in dc.OffersInBranch
on p.ID equals q.OfferID
where q.BranchID == singleLoc.LocationID
let value = (p.OriginalPrice - p.NewPrice) * 100 / p.OriginalPrice
orderby value descending
select new
{
Title = p.OfferTitle,
Description = p.Description,
BestOffer=value,
ID=p.ID,
LocationID=q.BranchID,
LocationName=q.CustomerBranch.BranchName,
OriginalPrice=SqlFunctions.StringConvert((decimal)p.OriginalPrice),
NewPrice=SqlFunctions.StringConvert((decimal)p.NewPrice),
StartDate=p.StartDate
})
.ToList()
.Select(x => new Offer()
{
Title = x.OfferTitle,
Description = x.Description,
BestOffer=value,
ID=x.ID,
LocationID=x.BranchID,
LocationName=x.CustomerBranch.BranchName,
OriginalPrice=x.OriginalPrice,
NewPrice=x.NewPrice,
StartDate=x.StartDate.ToShortDateString()
}).First();