Develop Reference

LINQ select more fields of same type

For simple example, I have a class like this:
public class MyClass {
public string name;
public int item1;
public int item2;
}
and a List(MyClass), how do I query the list to create List(int) of both fields item1 and item2? It seems it should be simple but I am struggling a long time.
var result = myList.Select(i => i.item1).ToList() //selects only one field
I know i could use anonymous type but since both item1 and item2 are integers I dont need any new type.
var result = myList.Select(i => new { i.item1, i.item2} ).ToList() //dont need new type, both are integers
how to create list of int? Or did I misunderstood what the anonymous types do?

If you don't like using anonymous types, you could always use a Tuple
var result = myList.Select(i => Tuple.Create(i.item1, i.item2) )
But since both item1 and item2 are integers, you can use an array:
var result = myList.Select(i => new[] { i.item1, i.item2 } )
This will result in a IEnumerable<int[]>. If you want an IEnumerable<int> (with each record's item1 and item2 together in one result set), use SelectMany:
var result = myList.SelectMany(i => new[] { i.item1, i.item2 } )

If you want a flattened list you can do:
List<int> ints = myList.SelectMany(i => new[] { i.item1, i.item2 }).ToList();
if you want to keep the values together you can create a tuple:
List<Tuple<int, int>> pairs = myList.Select(i => Tuple.Create(i.item1, i.item)).ToList():

You can use the following expression for example to create a list containing both items fro all elements of the original list myList.
var result = myList.Select(i => item1).Concat(myList.Select(i => i.item2));

How can I use GroupBy with a list of objects and return IList<T>?

I have a method that, must return a list of objects, but i don't know how!
my problem is converting(or maybe casting)groupdResult to IList.
internal IList<MyClass> MyMethod()
{
IList<MyClass> result=new List<MyClass>();
IList<MyClass> rawData = this.GetRawDatas();
foreach (MyClass item in rawDatas){
// do somthings
MyClass balancedData = new MyClass();
if(some conditions){
result.Add(balancedData);
}
}
var groupdResult = result.GroupBy(x => x.MyField).ToList();
return groupdResult;
}

Think about it: The result is either going to be a list of groups or a list of ungrouped objects. I believe that what you want is actually just a sorted list:
return result.OrderBy(x => x.MyField).ToList();

It's not the List/IList you're having trouble with. It's the fact that GroupBy does not return MyClass.
GroupBy returns an IGrouping, so the line
var groupdResult = result.GroupBy(x => x.MyField).ToList();
returns a List< IGrouping< MyClass > >.

Changing
var groupdResult = result.GroupBy(x => x.MyField).ToList();
to
var groupdResult = result.GroupBy(x => x.MyField).SelectMany(g=>g.ToList).ToList();
Solve the problem

Create List<int> from a list<class>

I have a class that is defined like this:
class A
{
int contractID;
string name;
string blah;
etc...
}
Multiple ClassA are stored stored in List(). I and to create a List based off all the ContractID's in List. Is there a way in LINQ to be able to do this instead of iterating through List()?

You can use the Select extension method to create a projection of the contractID. Like:
var listOfClass = new List<ClassA>(); //Or however you get your list.
var contractIDs = listOfClass.Select(x => x.contractID).ToList();

var onlyContractIDs = myList.Select(x => x.contractID).ToList()

List<ClassA> input = ...;
List<int> output = input.Select(a => a.contractID).ToList();
Select is a generic method. Explicitly specified, it would be Select<ClassA, int>, specifying the input and output types.
The ToList() at the end converts from an IEnumerable<int> to a List<int>, and it does the iteration over the input list now, rather than when the IEnumerable<int> is enumerated.

If you have a List you can make a List with LINQ like this
List<int> contractIDs = mycollection.Select(a => a.ContractID).ToList();

var list = new List<ClassA>();
...
list.Select(a => a.ContractID);

Merging List<> and mainting identity

I have 3 .net Lists items, I need to merge them all into one, to order them and bind them to a datagrid. However, I need a way of indicating which original list each item came from, so that I can identify this in the datagrid (change color or font etc).
Can anyone suggest the best way to do this?

List<Foo> list1 = new List<Foo>();
List<Foo> list2 = new List<Foo>();
List<Foo> list3 = new List<Foo>();
var query = list1.Select(foo => new { foo, list = "list1" })
.Concat(list2.Select(foo => new { foo, list = "list2" }))
.Concat(list3.Select(foo => new { foo, list = "list3" }))
.OrderBy(item => item.foo); // whatever you need to order by
Expand the properties as needed.

Assuming that your lists contains items of classes that you can amend I'd suggest that you add a property to those classes that keeps track of which type of the 3 it is. Either as an enum or possibly a reference to the actual list that contained it if you might need to refer back.
If you're not able to do that but assuming that they do contain a name property or similar and it's a readonly grid, a very ugly way would be to add a specific prefix/postfix to the name that says where it came from and then just remove that prefix/postfix before showing it on the screen.

Simple solution, assuming you don't want to modify the original class, or it's a primitive, you can use anonymous types:
var resultList = list1.Select(value => new {value, list = list1})
.Concat(list2.Select(value => new {value, list = list2})
.Concat(list3.Select(value => new {value, list = list3})))
.ToList();

I'd go with something like this:
List<object> list1 = new List<object>();
List<object> list2 = new List<object>();
List<object> list3 = new List<object>();
List<KeyValuePair<int, object>> mergedList = new List<KeyValuePair<int, object>>();
mergedList.AddRange(list1.Select(obj => new KeyValuePair<int, object>(1, obj)));
mergedList.AddRange(list2.Select(obj => new KeyValuePair<int, object>(2, obj)));
mergedList.AddRange(list3.Select(obj => new KeyValuePair<int, object>(3, obj)));
The better solution though, would be to add a property to your object that is some kind of enumeration that tells you something about the object itself. The lists themselves are metadata of some sort - list1 is a list of X, not just a list, so all of it's elements should have some kind of notion of X.

C# .NET 3.5 - Performing list like operations on members of a list

I have a field object and I create a list of fields:
class Field {
string objectName;
string objectType;
string fieldName;
string fieldValue;
//constructor...
}
List<Field> fieldList = new List<Field>();
Suppose I wanted to query this list to return a collection of distinct object names (to then be inserted into a checkedlistbox. How would I go about doing that?
I imagine some LINQ magic can manage this?

The expression should return a List of distinct object names from the list as defined. I converted it to a list since the docs for the CheckedListBox DataSource property indicated that it needs to implement IList or IListSource, not merely IEnumerable.
((ListControl)cbListBox).DataSource = fieldList.Select( f => f.objectName )
.Distinct()
.ToList() );
If accessing the checkedListBox as a ListControl doesn't give access to the DataSource (sometimes the docs lie), you could try:
cbListBox.Items.AddRange( fieldList.Select( f => f.objectName )
.Distinct()
.ToArray() );

Either of these work
Using var
1) var fieldNameCollection = from f in fieldList select f.FieldName;
2) Lambda syntax
var fieldNameCollection = fieldList.Select(f => f.FieldName);
Alternately, instead of using var, you can also use
IEnumerable fieldNameCollection = fieldList.Select(f => f.FieldName);

var q = from Field f in fileldList select f.objectName;
chkBoxList.DataSource = q.Distinct();