I am reading many sites to get a better idea of Linq -Group Join.
var customers = new Customer[]
{
new Customer{Code = 5, Name = "Sam"},
new Customer{Code = 6, Name = "Dave"},
new Customer{Code = 7, Name = "Julia"},
new Customer{Code = 8, Name = "Sue"}
};
// Example orders.
var orders = new Order[]
{
new Order{KeyCode = 5, Product = "Book"},
new Order{KeyCode = 6, Product = "Game"},
new Order{KeyCode = 7, Product = "Computer"},
new Order{KeyCode = 7, Product = "Mouse"},
new Order{KeyCode = 8, Product = "Shirt"},
new Order{KeyCode = 5, Product = "Underwear"}
};
var query = customers.GroupJoin(orders,
c => c.Code,
o => o.KeyCode,
(c, result) => new Result(c.Name, result));//why mention c here??
// Enumerate results.
foreach (var result in query)
{
Console.WriteLine("{0} bought...", result.Name);
foreach (var item in result.Collection)
{
Console.WriteLine(item.Product);
}
}
I couldnt understand why it gives (c, result) ? what if wrote as (c,o) ?
Can anyone share ideas on this?
These are just names of arguments passed to Func. You can use any name you want if that makes code more clear for you ie:
var query = customers.GroupJoin(orders,
c => c.Code,
o => o.KeyCode,
(something1, something2) => new Result(something1.Name, something2));
as it will just pass arguments from two previous Funcs into last one that is Func<TOuter, IEnumerable<TInner>, TResult>, so in that case Func<Customer, IEnumerable<Order>, Result>.
It's the same as with such situation:
public Result DoStuff(Order nameMeAnyWayYouWant, Customer meToo)
{
//do stuff here
}
Code from question is from: http://www.dotnetperls.com/groupjoin
I'm adding model classes that author skipped if anyone wants to elaborate and in case dotnetperls.com went down:
class Customer
{
public int Code { get; set; }
public string Name { get; set; }
}
class Order
{
public int KeyCode { get; set; }
public string Product { get; set; }
}
class Result
{
public string Name { get; set; }
public IEnumerable<Order> Collection { get; set; }
public Result(string name, IEnumerable<Order> collection)
{
this.Name = name;
his.Collection = collection;
}
}
Related
So here I have some code, which works ok. But I want to change the select part to something else, I am not sure what other methods I can use any help would be appreciated.
using System;
using System.Linq;
using System.Collections.Generic;
public class Program
{
public static void Main()
{
var all = new List<People>{new People{Id = 1, Name = "andy1", Age = null}, new People{Id = 2, Name = "andy2", Age = null}, new People{Id = 3, Name = "andy3", Age = null}, new People{Id = 4, Name = "andy4", Age = null}, };
var someOfThem = new List<People>{new People{Id = 1, Name = null, Age = 1}, new People{Id = 2, Name = null, Age = 1},new People{Id = 3, Name = null, Age = 1}};
var test = someOfThem.Select(c =>
{
c.Name = all.Find(a => a.Id == c.Id).Name;
return c;
});
foreach (var item in test)
Console.WriteLine("{0}={1}={2}", item.Id, item.Name, item.Age);
}
}
public class People
{
public int Id
{
get;
set;
}
public int? Age
{
get;
set;
}
public string Name
{
get;
set;
}
}
And here is the result.
1=andy1=1
2=andy2=1
3=andy3=1
I am just wondering is there another way to achieve the same result but a more elegant way? or an easier way?
var test = someOfThem.Select(c =>
{
c.Name = all.Find(a => a.Id == c.Id).Name;
return c;
});
Update
Sorry I did not show my problem properly at first, I have updated my quesiton. Please have a look again.
You can use C#'s LINQ keywords and more specifically, the join keyword assosciated with it:
using System;
using System.Linq;
using System.Collections.Generic;
public class Program
{
public static void Main()
{
var all = new List<People> { new People { Id = 1, Name = "andy1", }, new People { Id = 2, Name = "andy2", }, new People { Id = 3, Name = "andy3", }, new People { Id = 4, Name = "andy4", }, };
var someOfThem = new List<People> { new People { Id = 1, Name = null, }, new People { Id = 2, Name = null, } };
var test = from item in someOfThem
join element in all on item.Id equals element.Id
select element;
foreach (var item in test)
Console.WriteLine("{0}={1}", item.Id, item.Name);
}
}
public class People
{
public int Id
{
get;
set;
}
public string Name
{
get;
set;
}
}
The code version would be
var test = someOfThem.Join(all, item => item.Id, element => element.Id, (item, element) => element);
as shown in Robert's comment
You can use the Join (you can also use a dictionary, but I'm not going to show it):
Here's the syntax for join:
var test = someOfThem.Join(all, item => item.Id, element => element.Id,
(item, element) => new Person {
Id = item.Id ?? element.Id,
Name = item.Name ?? element.Name,
Age = item.Age ?? element.Age
});
You can implement Equals and GetHashCode in your People class and use Intersect.
Or, create an EqualityComparer, that way your comparison logic is decoupled:
class Program
{
public static void Main()
{
var all = new List<People> { new People { Id = 1, Name = "andy1", }, new People { Id = 2, Name = "andy2", }, new People { Id = 3, Name = "andy3", }, new People { Id = 4, Name = "andy4", }, };
var someOfThem = new List<People> { new People { Id = 1, Name = null, }, new People { Id = 2, Name = null, } };
var test = all.Intersect(someOfThem, new PeopleIdComparer()).ToList();
foreach (var item in test)
Console.WriteLine("{0}={1}", item.Id, item.Name);
}
}
public class PeopleIdComparer : IEqualityComparer<People>
{
public bool Equals(People x, People y)
{
return x.Id == y.Id;
}
public int GetHashCode(People obj)
{
return HashCode.Combine(obj.Id);
}
}
public class People
{
public int Id
{
get;
set;
}
public string Name
{
get;
set;
}
}
The best way I can describe what I'm trying to do is "Nested DistinctBy".
Let's say I have a collection of objects. Each object contains a collection of nicknames.
class Person
{
public string Name { get; set; }
public int Priority { get; set; }
public string[] Nicknames { get; set; }
}
public class Program
{
public static void Main()
{
var People = new List<Person>
{
new Person { Name = "Steve", Priority = 4, Nicknames = new string[] { "Stevo", "Lefty", "Slim" }},
new Person { Name = "Karen", Priority = 6, Nicknames = new string[] { "Kary", "Birdie", "Snookie" }},
new Person { Name = "Molly", Priority = 3, Nicknames = new string[] { "Mol", "Lefty", "Dixie" }},
new Person { Name = "Greg", Priority = 5, Nicknames = new string[] { "G-man", "Chubs", "Skippy" }}
};
}
}
I want to select all Persons but make sure nobody selected shares a nickname with another. Molly and Steve both share the nickname 'Lefty' so I want to filter one of them out. Only the one with highest priority should be included. If there is a highest priority tie between 2 or more then just pick the first one of them. So in this example I would want an IEnumerable of all people except Steve.
EDIT: Here's another example using music album instead of person, might make more sense.
class Album
{
string Name {get; set;}
int Priority {get;set;}
string[] Aliases {get; set;}
{
class Program
{
var NeilYoungAlbums = new List<Album>
{
new Person{ Name = "Harvest (Remastered)", Priority = 4, Aliases = new string[] { "Harvest (1972)", "Harvest (2012)"}},
new Person{ Name = "On The Beach", Priority = 6, Aliases = new string[] { "The Beach Album", "On The Beach (1974)"}},
new Person{ Name = "Harvest", Priority = 3, Aliases = new string[] { "Harvest (1972)"}},
new Person{ Name = "Freedom", Priority = 5, Aliases = new string[] { "Freedom (1989)"}}
};
}
The idea here is we want to show his discography but we want to skip quasi-duplicates.
I would solve this using a custom IEqualityComparer<T>:
class Person
{
public string Name { get; set; }
public int Priority { get; set; }
public string[] Nicknames { get; set; }
}
class PersonEqualityComparer : IEqualityComparer<Person>
{
public bool Equals(Person x, Person y)
{
if (x == null || y == null) return false;
return x.Nicknames.Any(i => y.Nicknames.Any(j => i == j));
}
// This is bad for performance, but if performance is not a
// concern, it allows for more readability of the LINQ below
// However you should check the Edit, if you want a truely
// LINQ only solution, without a wonky implementation of GetHashCode
public int GetHashCode(Person obj) => 0;
}
// ...
var people = new List<Person>
{
new Person { Name = "Steve", Priority = 4, Nicknames = new[] { "Stevo", "Lefty", "Slim" } },
new Person { Name = "Karen", Priority = 6, Nicknames = new[] { "Kary", "Birdie", "Snookie" } },
new Person { Name = "Molly", Priority = 3, Nicknames = new[] { "Mol", "Lefty", "Dixie" } },
new Person { Name = "Greg", Priority = 5, Nicknames = new[] { "G-man", "Chubs", "Skippy" } }
};
var distinctPeople = people.OrderBy(i => i.Priority).Distinct(new PersonEqualityComparer());
EDIT:
Just for completeness, this could be a possible LINQ only approach:
var personNicknames = people.SelectMany(person => person.Nicknames
.Select(nickname => new { person, nickname }));
var groupedPersonNicknames = personNicknames.GroupBy(i => i.nickname);
var duplicatePeople = groupedPersonNicknames.SelectMany(i =>
i.OrderBy(j => j.person.Priority)
.Skip(1).Select(j => j.person)
);
var distinctPeople = people.Except(duplicatePeople);
A LINQ-only solution
var dupeQuery = people
.SelectMany( p => p.Nicknames.Select( n => new { Nickname = n, Person = p } ) )
.ToLookup( e => e.Nickname, e => e.Person )
.SelectMany( e => e.OrderBy( p => p.Priority ).Skip( 1 ) );
var result = people.Except( dupeQuery ).ToList();
See .net fiddle sample
This works once, then you have to clear the set. Or store the results in a collection.
var uniqueNicknames = new HashSet<string>();
IEnumerable<Person> uniquePeople = people
.OrderBy(T => T.Priority) // ByDescending?
.Where(T => T.Nicknames.All(N => !uniqueNicknames.Contains(N)))
.Where(T => T.Nicknames.All(N => uniqueNicknames.Add(N)));
I am working on code which will give Cartesian product of two anonymous types. These 2 anonymous types are generated from database.
Code for 1st anonymous type:
private IEnumerable<object> GetItem()
{
return _unitOfWork.GetRepository<Item>()
.ListAll()
.Select(x => new
{
itemId = x.Id,
itemName = x.Name
})
}
Code for 2nd anonymous type:
private IEnumerable<object> GetVenue()
{
return _unitOfWork.GetRepository<Venue>()
.ListAll()
.Select(x => new
{
locationName = x.Address.City,
venueId = x.VenueId,
venueName = x.Name
})
}
I have following method to get the data and perform Cartesian product and return the data.
public object GetRestrictLookupInfo(IEnumerable<int> lookupCombinations)
{
IEnumerable<object> restrictList = new List<object>();
if (lookupCombinations.Contains(1))
{
var tempProductProfileList = GetItem();
restrictList = tempProductProfileList.AsEnumerable();
}
if (lookupCombinations.Contains(2))
{
var tempProductGroupList = GetVenue();
restrictList = (from a in restrictList.AsEnumerable()
from b in tempProductGroupList.AsEnumerable()
select new { a, b });
}
return restrictList;
}
I have controller which calls this method and return data in json format.
Controller Code
public HttpResponseMessage GetData(IEnumerable<int> lookupCombinations)
{
var lookupRestrictInfo = _sellerService.GetRestrictLookupInfo(lookupCombinations);
return Request.CreateResponse(HttpStatusCode.OK, lookupRestrictInfo);
}
Response expected is:-
[ {
"itemId": 1,
"itemName": "Music",
"locationName": "Paris",
"venueId": 99,
"venueName": "Royal Festival Hall"
} ]
Response which I receive is
[ {
"a": {
"itemId": 1,
"itemName": "Music"
},
"b": {
"locationName": "Paris",
"venueId": 99,
"venueName": "Royal Festival Hall" } }]
I am not able to get the expected JSON string.
You should start with the simplest possible code that shows your problem; your code above has a lot of complexities that may (or may not) have anything to do with your problem. Is this about manipulating anonymous types? Doing a Cartesian product with LINQ? Converting an object to JSON?
Here's one possible answer to what you might be looking for; notice that you can pass around anonymous types using generics instead of object.
namespace AnonymousTypes
{
class Program
{
static string Serialize(object o)
{
var d = (dynamic)o;
return d.ItemId.ToString() + d.ItemName + d.VenueId.ToString() + d.LocationName + d.VenueName;
}
static string GetData<T>(IEnumerable<T> result)
{
var retval = new StringBuilder();
foreach (var r in result)
retval.Append(Serialize(r));
return retval.ToString();
}
static string GetRestrictLookupInfo()
{
var restrictList = new[] { new { Id = 1, Name = "Music" }, new { Id = 2, Name = "TV" } };
var tempProductGroupList = new[] { new { LocationName = "Paris", Id = 99, Name = "Royal Festival Hall" } };
var result = from item in restrictList
from venue in tempProductGroupList
select new
{
ItemId = item.Id,
ItemName = item.Name,
LocationName = venue.LocationName,
VenueId = venue.Id,
VenueName = venue.Name
};
return GetData(result);
}
public static string GetData()
{
return GetRestrictLookupInfo();
}
static void Main(string[] args)
{
var result = GetData();
}
}
}
If that's not what you're looking for, you might start with code that doesn't use anonymous types, such as
namespace AnonymousTypes
{
sealed class Item
{
public int Id { get; set; }
public string Name { get; set; }
}
sealed class Venue
{
public string LocationName { get; set; }
public int Id { get; set; }
public string Name { get; set; }
}
sealed class ItemAndVenue
{
public int ItemId { get; set; }
public string ItemName { get; set; }
public string LocationName { get; set; }
public int VenueId { get; set; }
public string VenueName { get; set; }
}
class Program
{
static IEnumerable<Item> GetItem()
{
return new[] { new Item { Id = 1, Name = "Music" } };
}
static IEnumerable<Venue> GetVenue()
{
return new[] { new Venue { LocationName = "Paris", Id = 99, Name = "Royal Festival Hall" } };
}
static IEnumerable<ItemAndVenue> GetRestrictLookupInfo()
{
var restrictList = GetItem();
var tempProductGroupList = GetVenue();
var result = from item in restrictList
from venue in tempProductGroupList
select new ItemAndVenue
{
ItemId = item.Id,
ItemName = item.Name,
LocationName = venue.LocationName,
VenueId = venue.Id,
VenueName = venue.Name
};
return result;
}
static string GetData()
{
var v = GetRestrictLookupInfo().First();
return v.ItemId.ToString() + v.ItemName + v.VenueId.ToString() + v.LocationName + v.VenueName;
}
static void Main(string[] args)
{
var result = GetData();
}
}
}
In order to produce a single item in the output you need to create a new type, named or anonymous. Since you are using objects rather than actual types, the quickest approach is to cast them to dynamic:
var tempProductGroupList = GetVenue();
restrictList = (from a in restrictList.Cast<dynamic>()
from b in tempProductGroupList.Cast<dynamic>()
select new {
itemId = (int)a.itemId,
itemName = (string)a.itemName,
locationName = (string)b.locationName,
venueId = (int)b.venueId,
venueName = (string)b.venueName
});
This code is tightly coupled to the code producing both lists, because it assumes the knowledge of the field names of types passed into it dynamically. Any change in the structure of source data must be followed by a change in the code making combinations. In addition, it defeats run-time checking, so you need to be very careful with this code.
Try to create a simple object instead of nesting:
select new { a.itemId, a.itemName, b.locationName }
Like an option:
public object GetRestrictLookupInfo(IEnumerable<int> lookupCombinations)
{
List<Dictionary<string, object>> result = new List<Dictionary<string, object>>();
if (lookupCombinations.Contains(1))
{
var tmp = _unitOfWork.GetRepository<Item>()
.ListAll()
.Select(x => new
{
itemId = x.Id,
itemName = x.Name
})
.Select(x =>
{
var dic = new Dictionary<string, object>();
dic.Add(nameof(x.itemId), x.itemId);
dic.Add(nameof(x.itemName), x.itemName);
return dic;
});
result.AddRange(tmp);
}
if (lookupCombinations.Contains(2))
{
var tmp = _unitOfWork.GetRepository<Venue>()
.ListAll()
.Select(x => new
{
locationName = x.Address.City,
venueId = x.VenueId,
venueName = x.Name
})
.Select(x =>
{
var dic = new Dictionary<string, object>();
dic.Add(nameof(x.locationName), x.locationName);
dic.Add(nameof(x.venueId), x.venueId);
dic.Add(nameof(x.venueName), x.venueName);
return dic;
});
result = result.SelectMany(r => tmp.Select(t => r.Concat(t)));
}
return result;
}
It looks like some magic. I uses dictionary instead of object. It can be make in more clear way (extract few methods), but the idea should be clear.
Then, during serialization it will be presented as you need.
I have a Linq question: (DotNet Framework 4.0)
I have the following classes:
public class Employee
{
public Guid? EmployeeUUID { get; set; }
public string SSN { get; set; }
}
public class JobTitle
{
public Guid? JobTitleSurrogateKey { get; set; }
public string JobTitleName { get; set; }
}
public class EmployeeToJobTitleMatchLink
{
public EmployeeToJobTitleMatchLink()
{
this.TheJobTitle = new JobTitle() { JobTitleSurrogateKey = Guid.NewGuid(), JobTitleName = "SomeJobTitle:" + Guid.NewGuid().ToString("N") };
}
public Guid LinkSurrogateKey { get; set; }
/* Related Objects */
public Employee TheEmployee { get; set; }
public JobTitle TheJobTitle { get; set; }
}
public class Organization
{
public Organization()
{
this.Links = new List<EmployeeToJobTitleMatchLink>();
}
public int OrganizationSurrogateKey { get; set; }
public ICollection<EmployeeToJobTitleMatchLink> Links { get; set; }
}
In my code below, I can compare 2 child-collections and get the results I need (in "matches1".
Here I am using the "SSN" string property to compare and find the overlaps. And the Console.Write for matches1 works as I expect.
What I don't know how to do is compare the first child collection (org10) to all the children in (allOtherOrgsExceptOrg10 (all the Organizations and all the Links of these Organizations )
The commented out code shows kinda what I'm trying to do, one of my many feeble attempts today.
But basically, match2 would be populated with all the SSN overlaps...but comparing org10 with allOtherOrgsExceptOrg10, all their "Links", and their Employee.SSN's.
org10 overlaps with org20 with "AAA", so match2 would contain "AAA". and org10 overlaps with org30 with "BBB" so match2 would contain "BBB".
Organization org10 = new Organization();
org10.OrganizationSurrogateKey = 10;
Employee e11 = new Employee() { SSN = "AAA", EmployeeUUID = new Guid("AAAAAAAA-AAAA-AAAA-AAAA-AAAAAAAAAAAA") };
EmployeeToJobTitleMatchLink link11 = new EmployeeToJobTitleMatchLink();
link11.TheEmployee = e11;
org10.Links.Add(link11);
Employee e12 = new Employee() { SSN = "BBB", EmployeeUUID = new Guid("BBBBBBBB-BBBB-BBBB-BBBB-BBBBBBBBBBBB") };
EmployeeToJobTitleMatchLink link12 = new EmployeeToJobTitleMatchLink();
link12.TheEmployee = e12;
org10.Links.Add(link12);
Organization org20 = new Organization();
org20.OrganizationSurrogateKey = 20;
Employee e21 = new Employee() { SSN = "AAA", EmployeeUUID = new Guid("AAAAAAAA-AAAA-AAAA-AAAA-AAAAAAAAAAAA") };
EmployeeToJobTitleMatchLink link21 = new EmployeeToJobTitleMatchLink();
link21.TheEmployee = e21;
org20.Links.Add(link21);
Employee e22 = new Employee() { SSN = "CCC", EmployeeUUID = new Guid("CCCCCCCC-CCCC-CCCC-CCCC-CCCCCCCCCCCC") };
EmployeeToJobTitleMatchLink link22 = new EmployeeToJobTitleMatchLink();
link22.TheEmployee = e22;
org20.Links.Add(link22);
Organization org30 = new Organization();
org30.OrganizationSurrogateKey = 30;
Employee e31 = new Employee() { SSN = "BBB", EmployeeUUID = new Guid("BBBBBBBB-BBBB-BBBB-BBBB-BBBBBBBBBBBB") };
EmployeeToJobTitleMatchLink link31 = new EmployeeToJobTitleMatchLink();
link31.TheEmployee = e31;
org30.Links.Add(link31);
Employee e32 = new Employee();
e32.SSN = "ZZZ";
EmployeeToJobTitleMatchLink link32 = new EmployeeToJobTitleMatchLink();
link32.TheEmployee = e32;
org30.Links.Add(link32);
IList<Organization> allOtherOrgsExceptOrg10 = new List<Organization>();
/* Note, I did not add org10 here */
allOtherOrgsExceptOrg10.Add(org20);
allOtherOrgsExceptOrg10.Add(org30);
IEnumerable<EmployeeToJobTitleMatchLink> matches1 =
org10.Links.Where(org10Link => org20.Links.Any(org20Link => org20Link.TheEmployee.SSN.Equals(org10Link.TheEmployee.SSN, StringComparison.OrdinalIgnoreCase)));
IEnumerable<EmployeeToJobTitleMatchLink> matches2 = null;
//org10.Links.Where(org10Link => ( allOtherOrgs.Where ( anyOtherOrg => anyOtherOrg.Links.Any(dbSideChild => dbSideChild.TheEmployee.SSN == org10Link.TheEmployee.SSN)) );
if (null != matches1)
{
foreach (EmployeeToJobTitleMatchLink link in matches1)
{
Console.WriteLine(string.Format("matches1, SSN = {0}", link.TheEmployee.SSN));
}
}
if (null != matches2)
{
foreach (EmployeeToJobTitleMatchLink link in matches2)
{
Console.WriteLine(string.Format("matches2, SSN = {0}", link.TheEmployee.SSN));
}
}
matches2 =
allOtherOrgsExceptOrg10.SelectMany(x => x.Links)
.Where(x => org10.Links.Select(o => o.TheEmployee.SSN).Contains(x.TheEmployee.SSN));
You can use the SelectMany on the allOther collection to select all Links over all org's. Then check if any SSN is inside the org10 List.
See: http://msdn.microsoft.com/en-us/library/system.linq.enumerable.selectmany(v=vs.100).aspx
You can use SelectMany to flatten out the collection and then use it just like you have for matches1
IEnumerable<EmployeeToJobTitleMatchLink> matches2 =
org10.Links.Where(
org10Link =>
allOtherOrgsExceptOrg10.SelectMany(allOtherOrgs => allOtherOrgs.Links).Any(
anyOtherLink =>
anyOtherLink.TheEmployee.SSN.Equals(org10Link.TheEmployee.SSN, StringComparison.OrdinalIgnoreCase)));
The SelectMany will make it seem like one IEnumerable instead of and IEnumerable of an IEnumerable.
I got two classes, like:
public class Person
{
public long Id { get; set; }
public string Name { get; set; }
}
public class Vampire
{
public long Id { get; set; }
}
Then, I have two lists, a list of persons and a list of vampires. All vampires are persons.
What I need is two children lists of persons, infected and notInfected. I'm building the two lists with a for, but I know it's possible using linq or something.
Any help?
Something like this:
var vampireIds = new HashSet<long>(vampireList.Select(x => x.Id));
var infectedPersons = personList.Where(x => vampireIds.Contains(x.Id));
var regularPersons = personList.Where(x => !vampireIds.Contains(x.Id));
I would go with something like the following:
void Main()
{
var list = new List<Person>(){ new Person(){ Id = 1 }, new Vampire(){ Id = 2 } };
var infected = list.Where (x => x is Vampire);
var notInfected = list.Except(infected);
}
public class Person
{
public long Id { get; set; }
public string Name { get; set; }
}
public class Vampire : Person
{
}
If only a person can be a Vapire, you could inherit Vampire from Person and then iterate through all persons and see if they are Vampires; if yes -> add to Vampire list, otherwise to non-Vampire list.
Try this:
var people = new List<Person>
{
new Person {Id = 1, Name = "John"},
new Person {Name = "Dave", Id = 2},
new Person {Id = 3, Name = "Sarah"}
};
var vamps = new List<Vampire> {new Vampire {Id = 1}};
var theInfected = people.Where(p => vamps.Select(v => v.Id).Contains(p.Id));
var theAfraid = people.Except(theInfected);
foreach (var person in theInfected)
{
System.Console.WriteLine(person.Name + " Is Infected!");
}
foreach (var person in theAfraid)
{
System.Console.WriteLine(person.Name + " Is Afraid!");
}
Hope it's helpful.