緣起
最近在研讀 ASP.NET MVC 5 : 網站開發美學 Ch.03 時, 發現對於 SelectMany() 這個函式仍然不是很瞭解, 於是查了 MSDN 上的範例, 再加上配合 .NET Framework 源碼 一併進行 debug, 終於有了一些概念; 故撰寫本文, 以進行整理.完整範例, 請由此下載.
過程
說明相關程式檔案建立過程, 最後的版本, 可自行下載本範例. 在範例程式碼裡, 都有作了細節的說明, 可以參考.步驟1
自行建立 MyEnumerable 類別, 將 .NET Framework 源碼中 SelectMany() 相關擴充方法收錄至該類別, 以供除錯時, 得以明確看到程式執行過程.總共有 3 個 overloading 的擴充方法,
(1) 修改方法名稱, 在每個方法加上 My, 以與 .NET Framework 的源碼的原有方法作區隔
(2) 將有 Error.xxxx 的程式段註解掉,
註: .NET Framework 源碼 連結: ( http://referencesource.microsoft.com/#System.Core/System/Linq/Enumerable.cs,8f3471331178bcb0 )
Case 1:
SelectMany() 的參數說明:// 參數1: this IEnumerable<TSource> source --> 即 IEnumerable<PetOwner>
// 參數2: Func<TSource, IEnumerable<TResult>> selector --> 即 Func<PetOwner, IEnumerable<String>>
// source: PetOwner[] 陣列 (共 3 個元素)
// selector: petOwner => petOwner.Pets, 其中, petOwner 為 PetOwner 類別的物件實體
public static IEnumerable<TResult> MySelectMany<TSource, TResult>(this IEnumerable<TSource> source, Func<TSource, IEnumerable<TResult>> selector)
其實由 SelectMany() 實在看不太出來究竟回傳什麼, 由 SelectManyIterator() 來看, 就清楚多了.
foreach (TResult subElement in selector(element))
{
yield return subElement;
}
[註]: selector 為一個 delegate: Func<TSource, IEnumerable<TResult>>,
其傳入參數為 (element); 其實作為呼叫端的 petOwner => petOwner.Pets
#region CASE 1 的 SelectMany() 源碼
//參數1: this IEnumerable<TSource> source --> 即 IEnumerable<PetOwner>
//參數2: Func<TSource, IEnumerable<TResult>> selector --> 即 Func<PetOwner, IEnumerable<String>>
//source: PetOwner[] 陣列 (共 3 個元素)
//selector: petOwner => petOwner.Pets, 其中, petOwner 為 PetOwner 類別的物件實體
public static IEnumerable<TResult> MySelectMany<TSource, TResult>(this IEnumerable<TSource> source, Func<TSource, IEnumerable<TResult>> selector)
{
//if (source == null) throw Error.ArgumentNull("source");
//if (selector == null) throw Error.ArgumentNull("selector");
return MySelectManyIterator<TSource, TResult>(source, selector);
}
static IEnumerable<TResult> MySelectManyIterator<TSource, TResult>(IEnumerable<TSource> source, Func<TSource, IEnumerable<TResult>> selector)
{
//source: PetOwner[] 陣列 (共 3 個元素)
//element: PetOwner[] 陣列裡的每個元素, 即 PetOwner 的物件實體
//selector: petOwner => petOwner.Pets, 其中, petOwner 即為 element
foreach (TSource element in source)
{
//執行 selector(element) 之後, 會產出 petOwner.Pets 回傳
foreach (TResult subElement in selector(element))
{
yield return subElement;
}
}
}
#endregion
Case 2:
SelectMany() 的參數說明://參數1: this IEnumerable<TSource> source --> 即 IEnumerable<PetOwner>
//參數2: Func<TSource, int, IEnumerable<TResult>> selector --> 即 Func<PetOwner, int, IEnumerable<String>>
//source: PetOwner[] 陣列 (共 4 個元素)
//selector: (petOwner, index) => petOwner.Pets.Select(pet => index + pet), 其中, petOwner 為 PetOwner 類別的物件實體, index 為 int
public static IEnumerable<TResult> MySelectMany<TSource, TResult>(this IEnumerable<TSource> source, Func<TSource, int, IEnumerable<TResult>> selector)
foreach (TResult subElement in selector(element, index))
{
yield return subElement;
}
[註]: selector 為一個 delegate: Func<TSource, int, IEnumerable<TResult>>
其參數為 (element, index); 其實作為呼叫端的 (petOwner, index) => petOwner.Pets.Select(pet => index + pet)
#region CASE 2 的 SelectMany() 源碼
//參數1: this IEnumerable<TSource> source --> 即 IEnumerable<PetOwner>
//參數2: Func<TSource, int, IEnumerable<TResult>> selector --> 即 Func<PetOwner, int, IEnumerable<String>>
//source: PetOwner[] 陣列 (共 4 個元素)
//selector: (petOwner, index) => petOwner.Pets.Select(pet => index + pet), 其中, petOwner 為 PetOwner 類別的物件實體, index 為 int
public static IEnumerable<TResult> MySelectMany<TSource, TResult>(this IEnumerable<TSource> source, Func<TSource, int, IEnumerable<TResult>> selector)
{
// if (source == null) throw Error.ArgumentNull("source");
// if (selector == null) throw Error.ArgumentNull("selector");
return MySelectManyIterator<TSource, TResult>(source, selector);
}
static IEnumerable<TResult> MySelectManyIterator<TSource, TResult>(IEnumerable<TSource> source, Func<TSource, int, IEnumerable<TResult>> selector)
{
//source: PetOwner[] 陣列 (共 4 個元素)
//element: PetOwner[] 陣列裡的每個元素, 即 PetOwner 的物件實體
//selector: (petOwner, index) => petOwner.Pets.Select(pet => index + pet); 其中, petOwner 即為 element, index 為 int
// 註: index 代表陣列元素的位置, 由 0 起算
int index = -1;
foreach (TSource element in source)
{
checked { index++; }
foreach (TResult subElement in selector(element, index))
{
yield return subElement;
}
}
}
#endregion
Case 3:
SelectMany() 的參數說明://參數1: this IEnumerable<TSource> source --> 即 IEnumerable<PetOwner>
//參數2: Func<TSource, IEnumerable<TCollection>> collectionSelector --> 即 Func<PetOwner, IEnumerable<String>>
//參數3: Func<TSource, TCollection, TResult> resultSelector --> 即 Func<PetOwner, String, IEnumerable<String>>
//source: PetOwner[] 陣列 (共 4 個元素)
//collectionSelector: petOwner => petOwner.Pets
//resultSelector: (petOwner, petName) => new { petOwner, petName }
public static IEnumerable<TResult> MySelectMany<TSource, TCollection, TResult>(this IEnumerable<TSource> source, Func<TSource, IEnumerable<TCollection>> collectionSelector, Func<TSource, TCollection, TResult> resultSelector)
//把 element 傳入 collectionSelector 執行
//petOwner => petOwner.Pets ---> element => element.Pets
foreach (TCollection subElement in collectionSelector(element))
{
//把 element, subElement 傳入 resultSelector 執行
//(petOwner, petName) => new { petOwner, petName } ---> (element, subElement) => new { element, subElement }
yield return resultSelector(element, subElement);
}
[註]:
(1) collectionSelector 為一個 delegate: Func<TSource, IEnumerable<TCollection>>
其參數為 (element); 其實作為呼叫端的 petOwner => petOwner.Pets
(2) resultSelector 為一個 delegate: Func<TSource, TCollection, TResult>
其參數為 (element, subElement), 其實作為呼叫端的 (petOwner, petName) => new { petOwner, petName }
#region CASE 3 的 SelectMany() 源碼
//參數1: this IEnumerable<TSource> source --> 即 IEnumerable<PetOwner>
//參數2: Func<TSource, IEnumerable<TCollection>> collectionSelector --> 即 Func<PetOwner, IEnumerable<String>>
//參數3: Func<TSource, TCollection, TResult> resultSelector --> 即 Func<PetOwner, String, IEnumerable<String>>
//source: PetOwner[] 陣列 (共 4 個元素)
//collectionSelector: petOwner => petOwner.Pets
//resultSelector: (petOwner, petName) => new { petOwner, petName }
public static IEnumerable<TResult> MySelectMany<TSource, TCollection, TResult>(this IEnumerable<TSource> source, Func<TSource, IEnumerable<TCollection>> collectionSelector, Func<TSource, TCollection, TResult> resultSelector)
{
// if (source == null) throw Error.ArgumentNull("source");
// if (collectionSelector == null) throw Error.ArgumentNull("collectionSelector");
// if (resultSelector == null) throw Error.ArgumentNull("resultSelector");
return MySelectManyIterator<TSource, TCollection, TResult>(source, collectionSelector, resultSelector);
}
static IEnumerable<TResult> MySelectManyIterator<TSource, TCollection, TResult>(IEnumerable<TSource> source, Func<TSource, IEnumerable<TCollection>> collectionSelector, Func<TSource, TCollection, TResult> resultSelector)
{
//source: PetOwner[] 陣列 (共 4 個元素)
//element: PetOwner[] 陣列裡的每個元素, 即 PetOwner 的物件實體
//collectionSelector: petOwner => petOwner.Pets
//resultSelector: (petOwner, petName) => new { petOwner, petName }
foreach (TSource element in source)
{
//把 element 傳入 collectionSelector 執行
//petOwner => petOwner.Pets ---> element => element.Pets
foreach (TCollection subElement in collectionSelector(element))
{
//把 element, subElement 傳入 resultSelector 執行
//(petOwner, petName) => new { petOwner, petName } ---> (element, subElement) => new { element, subElement }
yield return resultSelector(element, subElement);
}
}
}
#endregion
步驟2
建立 PetOwner 類別class PetOwner
{
public string Name { get; set; }
public List<String> Pets { get; set; }
}
步驟3
在 Program.cs, 加入 3 個 static methods, 並修改 Main() 方法, 以呼叫那 3 個 methods(1) SelectManyEx1()
public static void SelectManyEx1()
{
PetOwner[] petOwners =
{ new PetOwner { Name="Higa, Sidney",
Pets = new List<string>{ "Scruffy", "Sam" } },
new PetOwner { Name="Ashkenazi, Ronen",
Pets = new List<string>{ "Walker", "Sugar" } },
new PetOwner { Name="Price, Vernette",
Pets = new List<string>{ "Scratches", "Diesel" } } };
// Query using SelectMany().
IEnumerable<string> query1 = petOwners.MySelectMany(petOwner => petOwner.Pets); // 呼叫自行由 .NET Framework 複製過來的擴充方法
//IEnumerable<string> query1 = petOwners.SelectMany(petOwner => petOwner.Pets); // 呼叫 .NET Framework 的擴充方法
Console.WriteLine("Using SelectMany():");
// Only one foreach loop is required to iterate
// through the results since it is a
// one-dimensional collection.
// 只需要 1 個 foreach 迴圈, 就可以選取第2層 Pets 的元素資料
foreach (string pet in query1)
{
Console.WriteLine(pet);
}
// This code shows how to use Select()
// instead of SelectMany().
IEnumerable<List<String>> query2 =
petOwners.Select(petOwner => petOwner.Pets);
Console.WriteLine("\nUsing Select():");
// Notice that two foreach loops are required to
// iterate through the results
// because the query returns a collection of arrays.
// 需要 2 個 foreach 迴圈
foreach (List<String> petList in query2)
{
foreach (string pet in petList)
{
Console.WriteLine(pet);
}
Console.WriteLine();
}
//輸出:
//Using SelectMany():
//Scruffy
//Sam
//Walker
//Sugar
//Scratches
//Diesel
//
//Using Select():
//Scruffy
//Sam
//
//Walker
//Sugar
//
//Scratches
//Diesel
}
(2) SelectManyEx2()
public static void SelectManyEx2()
{
PetOwner[] petOwners =
{ new PetOwner { Name="Higa, Sidney",
Pets = new List<string>{ "Scruffy", "Sam" } },
new PetOwner { Name="Ashkenazi, Ronen",
Pets = new List<string>{ "Walker", "Sugar" } },
new PetOwner { Name="Price, Vernette",
Pets = new List<string>{ "Scratches", "Diesel" } },
new PetOwner { Name="Hines, Patrick",
Pets = new List<string>{ "Dusty" } } };
// Project the items in the array by appending the index
// of each PetOwner to each pet's name in that petOwner's
// array of pets.
//
// IEnumerable<string> query =
// petOwners.SelectMany((petOwner, index) =>
// petOwner.Pets.Select(pet => index + pet));
//
IEnumerable<string> query =
petOwners.MySelectMany((petOwner, index) =>
petOwner.Pets.Select(pet => index + pet));
foreach (string pet in query)
{
Console.WriteLine(pet);
}
Console.WriteLine();
//輸出:
//
// 0Scruffy
// 0Sam
// 1Walker
// 1Sugar
// 2Scratches
// 2Diesel
// 3Dusty
}
(3) SelectManyEx3()
public static void SelectManyEx3()
{
PetOwner[] petOwners =
{ new PetOwner { Name="Higa",
Pets = new List<string>{ "Scruffy", "Sam" } },
new PetOwner { Name="Ashkenazi",
Pets = new List<string>{ "Walker", "Sugar" } },
new PetOwner { Name="Price",
Pets = new List<string>{ "Scratches", "Diesel" } },
new PetOwner { Name="Hines",
Pets = new List<string>{ "Dusty" } } };
// Project the pet owner's name and the pet's name.
//
//var query =
// petOwners
// .SelectMany(petOwner => petOwner.Pets, (petOwner, petName) => new { petOwner, petName })
// .Where(ownerAndPet => ownerAndPet.petName.StartsWith("S"))
// .Select(ownerAndPet =>
// new
// {
// Owner = ownerAndPet.petOwner.Name,
// Pet = ownerAndPet.petName
// }
// );
//
var query =
petOwners
.MySelectMany(petOwner => petOwner.Pets,
(petOwner, petName) => new { petOwner, petName }
)
.Select(ownerAndPet =>
new
{
Owner = ownerAndPet.petOwner.Name,
Pet = ownerAndPet.petName
}
);
//
//var query =
// petOwners
// .MySelectMany ( petOwner => petOwner.Pets,
// (petOwner, petName) => new { petOwner, petName }
// ) ;
//
// Print the results.
foreach (var obj in query)
{
Console.WriteLine(obj);
}
//本例輸出: (without Where() condition)
// {Owner=Higa, Pet=Scruffy}
// {Owner=Higa, Pet=Sam}
// {Owner=Ashkenazi, Pet=Walker}
// {Owner=Ashkenazi, Pet=Sugar}
// {Owner=Price, Pet=Scratches}
// {Owner=Price, Pet=Diesel}
// {Owner=Hines, Pet=Dusty}
//MSDN範例輸出: (with Where() condition)
// {Owner=Higa, Pet=Scruffy}
// {Owner=Higa, Pet=Sam}
// {Owner=Ashkenazi, Pet=Sugar}
// {Owner=Price, Pet=Scratches}
}
(4) Main()
static void Main(string[] args)
{
// CASE 1
Console.WriteLine("====== CASE 1 ======");
SelectManyEx1();
// CASE 2
Console.WriteLine("====== CASE 2 ======");
SelectManyEx2();
// CASE 3
Console.WriteLine("====== CASE 3 ======");
SelectManyEx3();
Console.ReadLine();
}
總結
要瞭解 SelectMany() 這個置於 Enumerable.cs 的擴充方法, 必須要有泛型及委派的基礎, 不然, 還真的不容易看懂.筆者學藝不精, 不確定 SelectMany() 在實務上用到的機會是否很大, 但可作為對泛型及委派觀念再次釐清之用.
MSDN 的這個範例, 可以應用在 Master Detail 結構的資料存取上, 例如: OrderMaster, OrderDetail,
參考文件
.
沒有留言:
張貼留言