LINQ 學習筆記（一）綜述（二）：.NET 1.x 到 2.0 的集合操作演進

.NET 1.x 時代：ArrayList 與裝箱/拆箱之痛

集合初始化

ArrayList list = new ArrayList();
list.Add(new Employee("Zxjay", 20, "010-123456"));
list.Add(new Employee("Andy", 30, "020-123456"));
list.Add(new Employee("Bill", 50, "010-3456789"));
list.Add(new Employee("Lee", 40, "0532-234567"));

所有元素視為 object，存入值型別需裝箱，取出需拆箱並強制轉型，效能損耗大且無編譯期型別安全。

排序：實作 IComparer

class SortArrayList : IComparer
{
    public int Compare(object x, object y)
    {
        Employee a = x as Employee;
        Employee b = y as Employee;
        return a.Name.CompareTo(b.Name);
    }
}
list.Sort(new SortArrayList());

篩選：手動迴圈與型別轉換

ArrayList selectedList = new ArrayList();
foreach (object obj in list)
{
    Employee emp = obj as Employee;
    if (emp != null && emp.Phone.StartsWith("010") && emp.Age < 50)
    {
        selectedList.Add(emp);
    }
}

輸出：重複的型別檢查與轉換

private void ShowList(ArrayList list)
{
    Console.WriteLine("{0,-20}{1,-5}{2}", "Name", "Age", "Phone");
    foreach (object obj in list)
    {
        Employee emp = obj as Employee;
        if (emp != null)
        {
            Console.WriteLine("{0,-20}{1,-5}{2}", emp.Name, emp.Age, emp.Phone);
        }
    }
}

.NET 2.0 時代：泛型與匿名方法的救贖

• 泛型：List<Employee> 移除裝箱/拆箱，編譯期強型別檢查
• 匿名方法：delegate(Employee x, Employee y) { ... } 簡化排序/篩選邏輯傳遞

排序與篩選：匿名方法

list.Sort(delegate(Employee x, Employee y) { return x.Name.CompareTo(y.Name); });

List<Employee> listBeijing = list.FindAll(
    delegate(Employee emp) { return emp.Phone.StartsWith("010") && emp.Age < 50; }
);

輸出：foreach 直接存取強型別元素

private void ShowList(IList<Employee> list)
{
    Console.WriteLine("{0,-20}{1,-5}{2}", "Name", "Age", "Phone");
    foreach (Employee emp in list)
    {
        Console.WriteLine("{0,-20}{1,-5}{2}", emp.Name, emp.Age, emp.Phone);
    }
}

演進總結

結語

從 ArrayList 到 List<T>，再到 LINQ，每一次演進都在解決前一代的痛點：型別安全、程式碼冗長、可讀性。理解這段歷史，有助於體會 LINQ 設計初衷與優雅之處。