Develop Reference

LINQ join failing between table and list of objects

I need to perform an update on a table with values from a List of objects in C# .NET Core 3.0. I tried to use the Join method, but receive this error:
Processing of the LINQ expression
DbSet<Room>
.Join(
outer: __p_0,
inner: p => p.RoomId,
outerKeySelector: s => s.ruId,
innerKeySelector: (s, p) => new {
kuku = s,
riku = p
})
by 'NavigationExpandingExpressionVisitor' failed. This may indicate either a bug or a limitation in EF Core. See link for more detailed information.
public class Room
{
[DatabaseGenerated(DatabaseGeneratedOption.None)]
[Key]
public int RoomId { get; set; }
[StringLength(50, MinimumLength = 3)]
public string RoomAddress { get; set; }
}
public class roomsForUpdate
{
public int ruId { get; set; }
public string ruName { get; set; }
}
var roomList = new List<roomsForUpdate>() { new roomsForUpdate { ruId = 1, ruName = "aa" }, new roomsForUpdate { ruId = 2, ruName = "bb" } };
var result = _context.Room.Join(roomList, p => p.RoomId, s => s.ruId, (s, p) => new { kuku = s, riku = p }).ToList();

You cannot join the EF Core LINQ query with a local list, because it can't be translated into SQL. Better first you get the database data and then join in memory.

LINQ is not meant to change the sources, it can only extract data from the sources. If you need to update data, you first fetch the items that must be updated, then you update them. Alternatively you can use plain old SQL to update the data without fetching it first.
In local memory, you have a sequence of RoomsForUpdate. Every RoomForUpdate has an Id (RuId) and a Name.
In your database you have a table with Rooms, Every Room in this table has an Id in RoomId and a RoomAddress.
It seems to me, that you want to update all Rooms that have an RoomId, that is one of the RuIds in your sequence of RoomsForUpdate. In other words: fetch (some properties of) all Rooms that have a value for RoomId that is a RuId in your sequence of RoomsForUpdate:
var roomsToUpdate = new List<roomsForUpdate>()
{
new roomsForUpdate { ruId = 1, ruName = "aa" },
new roomsForUpdate { ruId = 2, ruName = "bb" }
};
// Extract the Ids of the rooms that must be fetched
var roomToUpdateIds = roomsToUpdate.Select(room => room.ruId);
// Fetch all rooms from the database that have a RoomId that is in this sequence
var fetchedRooms = dbContext.Rooms
.Where(room => roomToUpdateIds.Contains(room => room.RoomId)
.ToList();
Of course you can put everything into one big LINQ statement. This will not improve efficiency, however it will deteriorate readability of your code.
Now to update the Rooms, you'll have to enumerate them one by one, and give the fetched rooms new values. You didn't say which new value you want. I have an inkling that you want to assign RuName to RoomAddress. This means that you have to combine the Room with the new value for the RoomAddress.
This can be done by LINQ:
var roomsWithExpectedNewValues = fetchedRooms.Join(roomsToUpdate,
fetchedRoom => fetchedRoom.RoomId, // from every fetched room take the Id
roomToUpdate => roomToUpdate.RuId, // from every room to update take the RuId
// for every fetchedRoom with its matching room to update, make one new:
(fetchedRoom, roomToUpdate) => new
{
Room = fetchedRoom,
NewValue = roomToUpdate.RuName,
})
.ToList();
To actually perform the update, you'll have to enumerate this sequence:
foreach (var itemToUpdate in roomsWithExpectedNewValues)
{
// assign RuName to RoomName
itemToUpdate.Room.RoomName = itemToUpdate.NewValue;
}
dbContext.SaveChanges();
A little less LINQ
Although this works, there seems to be a lot of magic going on. The join will internally make a Dictionary for fast lookup, and throws it away. I think a little less LINQ will make it way more easy to understand what's going on.
// your original roomsToUpdate
var roomsToUpdate = new List<roomsForUpdate>()
{
new roomsForUpdate { ruId = 1, ruName = "aa" },
new roomsForUpdate { ruId = 2, ruName = "bb" }
};
var updateDictionary = roomsToUpdate.ToDictionary(
room => room.RuId, // key
room => room.RuName) // value
The Keys of the dictionary are the IDs of the rooms that you want to fetch:
// fetch the rooms that must be updated:
var fetchedRooms = dbContext.Rooms
.Where(room => updateDictionary.Keys.Contains(room => room.RoomId)
.ToList();
// Update:
foreach (var fetchedRoom in fetchedRooms)
{
// from the dictionary fetch the ruName:
var ruName = updateDicationary[fetchedRoom.RoomId];
// assign the ruName to RoomAddress
fetchedRoom.RoomAddress = ruName;
// or if you want, do this in one statement:
fetchedRoom.RoomAddress = updateDicationary[fetchedRoom.RoomId];
}
dbContext.SaveChanges();

Inner join two lists based on a child list

I have two lists of different types of custom objects, and I'm trying to perform an inner join where the join criteria is contained within a child list of the objects.
Here's an example of my objects:
public class Container
{
public List<Reference> Refs;
public Container(List<Reference> refs )
{
Refs = refs;
}
}
public class Reference
{
public int Id;
public Reference(int id )
{
Id = id;
}
}
And here's an example the data I'm working with:
List<Container> containers = new List<Container>()
{
new Container(new List<Reference>()
{
new Reference(1),
new Reference(2),
new Reference(3)
}),
new Container(new List<Reference>()
{
new Reference(4),
new Reference(5),
new Reference(6)
})
};
List<Reference> references = new List<Reference>()
{
new Reference(4),
new Reference(5),
new Reference(6)
};
I'm trying to select all the Containers in List<Container> which have a matching Reference in the List<Reference> based on Reference.Id. With this data, I expect only the second item in the List<Container> to be selected.
If it were valid syntax, I'd be looking to do something along the lines of:
var query = from c in containers
join r in references on c.Refs.Contains( r.Id )
select c;
How can this be done? Thanks
Sorry for the poor title. I'm struggling to put this scenario into a short group of words - please suggest an edit if you can think of something more suitable. Thanks

an inner join is not necessary here, you're better off without it:
containers.Where(c => c.Refs.Any(x => references.Any(e => x.Id == e.Id)));
or if you want the entire set of Id's to be equal then use SequenceEqual:
var sequence = references.Select(e => e.Id);
var result = containers.Where(c => c.Refs.Select(s => s.Id).SequenceEqual(sequence));

containers.Where(c => c.Refs.Select(r => r.Id).Intersect(references.Select(r => r.Id)).Any());

I would use:
var query = from c in containers where c.Refs.SequenceEqual(references)
select c;
No join is necessary.

LINQ Select distinct on different variables

I have a query which selects a list of my class. It looks like so:
IQueryable<ClaimsBySupplierAggregate> agg =
(from d in alliance.SupplierSearchByReviewPeriod
where d.ClientID == ClientID && ReviewPeriodIDs.Contains((int)d.ReviewPeriodID)
select new ClaimsBySupplierAggregate {
Amount = d.Amount,
StatusCategoryID = d.StatusCategoryID,
DeptName = d.DepartmentName,
APLReason = d.APLReason,
Area = d.AreaDesc,
StatusCategoryDesc = d.StatusCategoryDesc,
Agreed = d.Agreed
});
Later on in the application I select each variable and get the distinct values like this:
SupplierModel.APLReason = agg.Select(r => r.APLReason).Distinct().ToList();
SupplierModel.AreaDesc = agg.Select(r => r.Area).Distinct().ToList();
SupplierModel.DeptName = agg.Select(r => r.DeptName).Distinct().ToList();
SupplierModel.StatCatDes = aggg.Select(r => r.StatusCategoryDesc).Distinct().ToList();
Is there a way to do this in one LINQ statement?

You could using Aggregate, but you would need a complex object for the seed, which brings you to the same complexity of code. I think that for this particular case, using LInQ enters the Golden Hammer antipattern. Just use an old fashioned loop and four HashSets instead of Lists and you are done and the code is more readable and your intention clearer.

I think you can use this then customize the code follow your expectation.
Or you can use group by.
public class LinqTest
{
public int id { get; set; }
public string value { get; set; }
public override bool Equals(object obj)
{
LinqTest obj2 = obj as LinqTest;
if (obj2 == null) return false;
return id == obj2.id;
}
public override int GetHashCode()
{
return id;
}
}
List<LinqTest> uniqueIDs = myList.Distinct().ToList();

There is a contrived way to do this:
from d in alliance.SupplierSearchByReviewPeriod
where d.ClientID == ClientID && ReviewPeriodIDs.Contains((int)d.ReviewPeriodID)
group d by 0 into g
select new
{
APLReasons = g.Select(d => d.APLReason).Distinct(),
AreaDescs = g.Select(d => d.Area).Distinct(),
DeptNames = g.Select(d => d.DeptName).Distinct(),
StatusCategoryDescs = g.Select(d => d.StatusCategoryDesc).Distinct(),
}
The part group d by 0 into g creates one group of all items, from which you can subsequently query any aggregate you want.
BUT...
...this creates a very inefficient query with UNIONs and (of course) DISTINCTs. It's probably better (performance-wise) to simply get the flat data from the database and do the aggregations in subsequent code.

How to aggregate duplicate records and sum value using C#

I want to create a list where it holds several agents and the number of calls they make and did it like so:
public class Agent
{
public string Agent_ID{ get; set; }
public string Name { get; set; }
public int Calls { get; set; }
}
var list = new List<Agent>() // To create a list to hold the data
{
new Agent() { Agent_ID = "TK_J", Name = "James", Calls = 10 },
new Agent() { Agent_ID = "TK_K", Name = "Kurtis", Calls = 10 },
new Agent() { Agent_ID = "TK_R", Name = "Rebecca", Calls = 5 },
new Agent() { Agent_ID = "TK_J", Name = "James", Calls = 10 },
new Agent() { Agent_ID = "TK_R", Name = "Rebecca", Calls = 5 },
new Agents(){ Agent_ID = "TK_B", Name = "Bobby", Calls = 10 },
};
As you can see there will be redundant lines of data. So I want to use C# aggregation function such as group by to sum up the similar agent's number of calls. What I was trying was:
list.GroupBy(i => i.Agent_ID).Select(g => new
{
Agent_ID= g.Key,
Name = */ How do i bring the name here*/,
Calls = g.Sum(i => i.Calls)});
Anyone can help me? Appreciate any help or advice. If there is something wrong teach me how to fix the code. Many thanks!!

You are currently grouping by only AgentID.
You'll need to project the fields you require as an anonymous object so they can be available as an IGrouping<annonymous,Agent> to the select.
See below.
list.GroupBy(i => new {i.Agent_ID, i.Name}).Select(g => new
{
Agent_ID= g.Key.Agent_ID,
Name = g.Key.Name,
Calls = g.Sum(i => i.Calls)
});

Assuming that Agent_ID will be synchronized with the Agent's Name property, you can also use an aggregate like First() to return the the name from any Agent record in the group, since all agents in each group will have the same name:
list.GroupBy(i => i.Agent_ID)
.Select(g => new Agent // Strong class
{
Agent_ID= g.Key.Agent_ID,
Name = g.First().Name,
Calls = g.Sum(i => i.Calls)
})
Also, since you seem to be projecting back into the same shape, why not retain and project into the strong class, rather than a new anon class?

Join table with object list

I have a table, lets say tblCar with all the related columns like Id, Make, Model, Color etc.
I have a search model for car containing two params Id and Model.
public class CarSearch
{
public int Id { get; set; }
public string Model { get; set; }
}
var carSearchObjets = new List<CarSearch>();
With list of primitive data (like Id list), to get cars with those Ids I could have done:
var idList = new List<int> { 1, 2, 3 };
var carsFromQuery = context.Cars.Where(x => idList.Contains(x.Id);
But if I have to fetch all the cars with Id and model from the list, how do I do it? Simple join cannot be done between in memory objects and tables.
I need something like,
from m in context.Cars
join n in carSearchObjets
on new { Id = n.Id, Model = n.Model } equals new { Id = m.Id, Model = m.Model }
select m;
This obviously won't work.
Please ignore any typos.And if you need more info or the question is not clear, let me know.

One (ugly-but-working) way to manage that is to use concatenation with a "never used" concat char.
I mean a char that should never appear in the datas. This is always dangerous, as... never is never sure, but you've got the idea.
For example, we'll say that our "never used" concat char will be ~
This is not good for perf, but at least working :
var carSearchObjectsConcatenated = carSearchObjets.Select(m => new { m.Id + "~" + m.Model});
then you can use Contains again (concatenating on the db too) : you'll need to use SqlFunctions.StringConvert if you wanna concatenate string and numbers on the db side.
var result = context.Cars.Where(m =>
carSearchObjectsConcatenated.Contains(SqlFunctions.StringConvert((double)m.Id) + "~" + m.Model);
EDIT
Another solution would be to use PredicateBuilder, as mentionned by Sorax, or to build your own Filter method if you don't want a third party lib (but PredicateBuilder is really fine).
Something like that in a static class :
public static IQueryable<Car> FilterCars(this IQueryable<Car> cars, IEnumerable<SearchCar> searchCars)
{
var parameter = Expression.Parameter(typeof (Car), "m");
var idExpression = Expression.Property(parameter, "Id");
var modelExpression = Expression.Property(parameter, "Model");
Expression body = null;
foreach (var search in searchCars)
{
var idConstant = Expression.Constant(search.Id);
var modelConstant = Expression.Constant(search.Model);
Expression innerExpression = Expression.AndAlso(Expression.Equal(idExpression, idConstant), Expression.Equal(modelExpression, modelConstant));
body = body == null
? innerExpression
: Expression.OrElse(body, innerExpression);
}
var lambda = Expression.Lambda<Func<Car, bool>>(body, new[] {parameter});
return cars.Where(lambda);
}
usage
var result = context.Cars.FilterCars(carSearchObjets);
this will generate an sql looking like
select ...
from Car
where
(Id = 1 And Model = "ax") or
(Id = 2 And Model = "az") or
(Id = 3 And Model = "ft")

'PredicateBuilder' might be helpful.
var predicate = PredicateBuilder.False<Car>();
carSearchObjects
.ForEach(a => predicate = predicate.Or(p => p.Id == a.Id && p.Model == a.Model));
var carsFromQuery = context.Cars.AsExpandable().Where(predicate);
Note the text in the link regarding EF:
If you're using Entity Framework, you'll need the complete LINQKit -
for the AsExpandable functionality. You can either reference
LINQKit.dll or copy LINQKit's source code into your application.

Old school solution..
//in case you have a
List<CarSearch> search_list; //already filled
List<Cars> cars_found = new List<Cars>();
foreach(CarSearch carSearch in search_list)
{
List<Cars> carsFromQuery = context.Cars.Where(x => x.Id == carSearch.Id && x.Model == carSearch.Model).ToList();
cars_found.AddRange(carsFromQuery);
}
Abd don't worry about the for loops.

I landed up passing in an xml list as a parameter to the sql query and joined to that:
var xml = new XElement("Cars", yourlist.Select(i => new XElement("Car", new XElement("Id", i.Id), new XElement("Model", i.Model))));
var results = Cars
.FromSql("SELECT cars.*"
+ "FROM #xml.nodes('/Cars/Car') Nodes(Node)"
+ "JOIN Cars cars on cars.Id = Nodes.Node.value('Id[1]', 'int') and cars.Model = Nodes.Node.value('Model[1]', 'varchar(100)')",
new SqlParameter("#xml", new SqlXml(xml.CreateReader())));
For entity-framework-core users I created a nuget package extension:
EntityFrameworkCore.SqlServer.Extensions.Contains