Linq的Distinct太不给力了[转]

假设我们有一个类：Product

public class Product
{
    public string Id { get; set; }
    public string Name { get; set; }
}

Main函数如下：

static void Main()
{
    List<Product> products = new List<Product>()
    {
        new Product(){ Id="1", Name="n1"},
        new Product(){ Id="1", Name="n2"},
        new Product(){ Id="2", Name="n1"},
        new Product(){ Id="2", Name="n2"},
    };

    var distinctProduct = products.Distinct();

    Console.ReadLine();
}

.csharpcode, .csharpcode pre
{
font-size: small;
color: black;
font-family: consolas, "Courier New", courier, monospace;
background-color: #ffffff;
/*white-space: pre;*/
}
.csharpcode pre { margin: 0em; }
.csharpcode .rem { color: #008000; }
.csharpcode .kwrd { color: #0000ff; }
.csharpcode .str { color: #006080; }
.csharpcode .op { color: #0000c0; }
.csharpcode .preproc { color: #cc6633; }
.csharpcode .asp { background-color: #ffff00; }
.csharpcode .html { color: #800000; }
.csharpcode .attr { color: #ff0000; }
.csharpcode .alt
{
background-color: #f4f4f4;
width: 100%;
margin: 0em;
}
.csharpcode .lnum { color: #606060; }

可以看到distinctProduct 的结果是：

因为Distinct 默认比较的是Product对象的引用，所以返回4条数据。

那么如果我们希望返回Id唯一的product，那么该如何做呢？

Distinct方法还有另一个重载：

//通过使用指定的 System.Collections.Generic.IEqualityComparer<T> 对值进行比较
//返回序列中的非重复元素。
 public static IEnumerable<TSource> Distinct<TSource>(this IEnumerable<TSource> source, 

IEqualityComparer<TSource> comparer);

该重载接收一个IEqualityComparer的参数。

假设要按Id来筛选，那么应该新建类ProductIdComparer 内容如下：

public class ProductIdComparer : IEqualityComparer<Product>
{
    public bool Equals(Product x, Product y)
    {
        if (x == null)
            return y == null;
        return x.Id == y.Id;
    }

    public int GetHashCode(Product obj)
    {
        if (obj == null)
            return 0;
        return obj.Id.GetHashCode();
    }
}

.csharpcode, .csharpcode pre
{
font-size: small;
color: black;
font-family: consolas, "Courier New", courier, monospace;
background-color: #ffffff;
/*white-space: pre;*/
}
.csharpcode pre { margin: 0em; }
.csharpcode .rem { color: #008000; }
.csharpcode .kwrd { color: #0000ff; }
.csharpcode .str { color: #006080; }
.csharpcode .op { color: #0000c0; }
.csharpcode .preproc { color: #cc6633; }
.csharpcode .asp { background-color: #ffff00; }
.csharpcode .html { color: #800000; }
.csharpcode .attr { color: #ff0000; }
.csharpcode .alt
{
background-color: #f4f4f4;
width: 100%;
margin: 0em;
}
.csharpcode .lnum { color: #606060; }

使用的时候，只需要

var distinctProduct = products.Distinct(new ProductIdComparer());

结果如下：

现在假设我们要 按照 Name来筛选重复呢？

很明显，需要再添加一个类ProductNameComparer.

那能不能使用泛型类呢？？

新建类PropertyComparer<T> 继承IEqualityComparer<T> 内容如下：

.csharpcode, .csharpcode pre
{
font-size: small;
color: black;
font-family: consolas, "Courier New", courier, monospace;
background-color: #ffffff;
/*white-space: pre;*/
}
.csharpcode pre { margin: 0em; }
.csharpcode .rem { color: #008000; }
.csharpcode .kwrd { color: #0000ff; }
.csharpcode .str { color: #006080; }
.csharpcode .op { color: #0000c0; }
.csharpcode .preproc { color: #cc6633; }
.csharpcode .asp { background-color: #ffff00; }
.csharpcode .html { color: #800000; }
.csharpcode .attr { color: #ff0000; }
.csharpcode .alt
{
background-color: #f4f4f4;
width: 100%;
margin: 0em;
}
.csharpcode .lnum { color: #606060; }

public class PropertyComparer<T> : IEqualityComparer<T>
{
    private PropertyInfo _PropertyInfo;

    /// <summary>
    /// 通过propertyName 获取PropertyInfo对象    

    /// </summary>
    /// <param name="propertyName"></param>
    public PropertyComparer(string propertyName)
    {
        _PropertyInfo = typeof(T).GetProperty(propertyName,
        BindingFlags.GetProperty | BindingFlags.Instance | BindingFlags.Public);
        if (_PropertyInfo == null)
        {
            throw new ArgumentException(string.Format("{0} is not a property of type {1}.",
                propertyName, typeof(T)));
        }
    }

    #region IEqualityComparer<T> Members

    public bool Equals(T x, T y)
    {
        object xValue = _PropertyInfo.GetValue(x, null);
        object yValue = _PropertyInfo.GetValue(y, null);

        if (xValue == null)
            return yValue == null;

        return xValue.Equals(yValue);
    }

    public int GetHashCode(T obj)
    {
        object propertyValue = _PropertyInfo.GetValue(obj, null);

        if (propertyValue == null)
            return 0;
        else
            return propertyValue.GetHashCode();
    }

    #endregion
}

.csharpcode, .csharpcode pre
{
font-size: small;
color: black;
font-family: consolas, "Courier New", courier, monospace;
background-color: #ffffff;
/*white-space: pre;*/
}
.csharpcode pre { margin: 0em; }
.csharpcode .rem { color: #008000; }
.csharpcode .kwrd { color: #0000ff; }
.csharpcode .str { color: #006080; }
.csharpcode .op { color: #0000c0; }
.csharpcode .preproc { color: #cc6633; }
.csharpcode .asp { background-color: #ffff00; }
.csharpcode .html { color: #800000; }
.csharpcode .attr { color: #ff0000; }
.csharpcode .alt
{
background-color: #f4f4f4;
width: 100%;
margin: 0em;
}
.csharpcode .lnum { color: #606060; }

主要是重写的Equals 和GetHashCode 使用了属性的值比较。

使用的时候，只需要：

//var distinctProduct = products.Distinct(new PropertyComparer<Product>("Id"));
var distinctProduct = products.Distinct(new PropertyComparer<Product>("Name"));

结果如下：
.csharpcode, .csharpcode pre
{
font-size: small;
color: black;
font-family: consolas, "Courier New", courier, monospace;
background-color: #ffffff;
/*white-space: pre;*/
}
.csharpcode pre { margin: 0em; }
.csharpcode .rem { color: #008000; }
.csharpcode .kwrd { color: #0000ff; }
.csharpcode .str { color: #006080; }
.csharpcode .op { color: #0000c0; }
.csharpcode .preproc { color: #cc6633; }
.csharpcode .asp { background-color: #ffff00; }
.csharpcode .html { color: #800000; }
.csharpcode .attr { color: #ff0000; }
.csharpcode .alt
{
background-color: #f4f4f4;
width: 100%;
margin: 0em;
}
.csharpcode .lnum { color: #606060; }

为什么微软不提供PropertyEquality<T> 这个类呢？

按照上面的逻辑，这个类应该没有很复杂啊，细心的同学可以发现PropertyEquality 大量的使用了反射。每次获取属性的值的时候，都在调用   
_PropertyInfo.GetValue(x, null);

可想而知，如果要筛选的记录非常多的话，那么性能无疑会受到影响。

为了提升性能，可以使用表达式树将反射调用改为委托调用，

具体代码如下：

public class FastPropertyComparer<T> : IEqualityComparer<T>
{
    private Func<T, Object> getPropertyValueFunc = null;

    /// <summary>
    /// 通过propertyName 获取PropertyInfo对象
    /// </summary>
    /// <param name="propertyName"></param>
    public FastPropertyComparer(string propertyName)
    {
        PropertyInfo _PropertyInfo = typeof(T).GetProperty(propertyName,
        BindingFlags.GetProperty | BindingFlags.Instance | BindingFlags.Public);
        if (_PropertyInfo == null)
        {
            throw new ArgumentException(string.Format("{0} is not a property of type {1}.",
                propertyName, typeof(T)));
        }

        ParameterExpression expPara = Expression.Parameter(typeof(T), "obj");
        MemberExpression me = Expression.Property(expPara, _PropertyInfo);
        getPropertyValueFunc = Expression.Lambda<Func<T, object>>(me, expPara).Compile();
    }

    #region IEqualityComparer<T> Members

    public bool Equals(T x, T y)
    {
        object xValue = getPropertyValueFunc(x);
        object yValue = getPropertyValueFunc(y);

        if (xValue == null)
            return yValue == null;

        return xValue.Equals(yValue);
    }

    public int GetHashCode(T obj)
    {
        object propertyValue = getPropertyValueFunc(obj);

        if (propertyValue == null)
            return 0;
        else
            return propertyValue.GetHashCode();
    }

    #endregion
}

可以看到现在获取值只需要getPropertyValueFunc(obj) 就可以了。

使用的时候：

var distinctProduct = products.Distinct(new FastPropertyComparer<Product>("Id")).ToList();