LINQ to SQL的版本GROUP BY与ROLLUP(LINQ to SQL version

2019-07-03 22:26发布

我试图重写一些老SQL到LINQ到SQL。 我有一组存储过程与ROLLUP,但我不知道LINQ相当于是什么。 LINQ拥有的GroupBy,但它并不像它支持汇总。

结果,我试图让会是这样的一个简单的例子:

+-----------+---------------+--------------------+
|   City    |  ServicePlan  |  NumberOfCustomers |
+-----------+---------------+--------------------+
| Seattle   |  Plan A       |                 10 |
| Seattle   |  Plan B       |                  5 |
| Seattle   |  All          |                 15 |
| Portland  |  Plan A       |                 20 |
| Portland  |  Plan C       |                 10 |
| Portland  |  All          |                 30 |
| All       |  All          |                 45 |
+-----------+---------------+--------------------+

关于我如何能使用LINQ to SQL获得这些结果的任何想法?

Answer 1:

我想出了一个更简单的解决方案。 我试图使它的方式复杂得多,它需要的是。 而不需要3-5类/方法我只需要一个方法。

基本上,你做你的排序和分组你自己,然后调用WithRollup()得到一个List<>与小计和总计的项目。 我无法弄清楚如何生成SQL一边因此那些与LINQ做对象的小计和总计。 下面的代码:

/// <summary>
/// Adds sub-totals to a list of items, along with a grand total for the whole list.
/// </summary>
/// <param name="elements">Group and/or sort this yourself before calling WithRollup.</param>
/// <param name="primaryKeyOfElement">Given a TElement, return the property that you want sub-totals for.</param>
/// <param name="calculateSubTotalElement">Given a group of elements, return a TElement that represents the sub-total.</param>
/// <param name="grandTotalElement">A TElement that represents the grand total.</param>
public static List<TElement> WithRollup<TElement, TKey>(this IEnumerable<TElement> elements,
    Func<TElement, TKey> primaryKeyOfElement,
    Func<IGrouping<TKey, TElement>, TElement> calculateSubTotalElement,
    TElement grandTotalElement)
{
    // Create a new list the items, subtotals, and the grand total.
    List<TElement> results = new List<TElement>();
    var lookup = elements.ToLookup(primaryKeyOfElement);
    foreach (var group in lookup)
    {
        // Add items in the current group
        results.AddRange(group);
        // Add subTotal for current group
        results.Add(calculateSubTotalElement(group));
    }
    // Add grand total
    results.Add(grandTotalElement);

    return results;
}

以及如何使用它的一个例子:

class Program
{
    static void Main(string[] args)
    {
        IQueryable<CustomObject> dataItems = (new[]
        {
            new CustomObject { City = "Seattle", Plan = "Plan B", Charges = 20 },
            new CustomObject { City = "Seattle", Plan = "Plan A", Charges = 10 },
            new CustomObject { City = "Seattle", Plan = "Plan B", Charges = 20 },
            new CustomObject { City = "Seattle", Plan = "Plan A", Charges = 10 },
            new CustomObject { City = "Seattle", Plan = "Plan A", Charges = 10 },
            new CustomObject { City = "Seattle", Plan = "Plan A", Charges = 10 },
            new CustomObject { City = "Portland", Plan = "Plan A", Charges = 10 },
            new CustomObject { City = "Portland", Plan = "Plan A", Charges = 10 },
            new CustomObject { City = "Portland", Plan = "Plan C", Charges = 30 },
            new CustomObject { City = "Portland", Plan = "Plan C", Charges = 30 },
            new CustomObject { City = "Portland", Plan = "Plan C", Charges = 30 }
        }).AsQueryable();

        IQueryable<CustomObject> orderedElements = from item in dataItems
                                                   orderby item.City, item.Plan
                                                   group item by new { item.City, item.Plan } into grouping
                                                   select new CustomObject
                                                   {
                                                       City = grouping.Key.City,
                                                       Plan = grouping.Key.Plan,
                                                       Charges = grouping.Sum(item => item.Charges),
                                                       Count = grouping.Count()
                                                   };

        List<CustomObject> results = orderedElements.WithRollup(
            item => item.City,
            group => new CustomObject
            {
                City = group.Key,
                Plan = "All",
                Charges = group.Sum(item => item.Charges),
                Count = group.Sum(item => item.Count)
            },
            new CustomObject
            {
                City = "All",
                Plan = "All",
                Charges = orderedElements.Sum(item => item.Charges),
                Count = orderedElements.Sum(item => item.Count)
            });

        foreach (var result in results)
            Console.WriteLine(result);

        Console.Read();
    }
}

class CustomObject
{
    public string City { get; set; }
    public string Plan { get; set; }
    public int Count { get; set; }
    public decimal Charges { get; set; }

    public override string ToString()
    {
        return String.Format("{0} - {1} ({2} - {3})", City, Plan, Count, Charges);
    }
}


Answer 2:

我知道了! 一个通用GroupByWithRollup。 它只有两列组,但可以很容易地扩展,以支持更多。 我也许会做另一种接受三列的版本。 键类/方法的分组<>,GroupByMany <>(),和GroupByWithRollup <>()。 小计()和GrandTotal()方法是助手当你实际使用GroupByWithRollup <>()。 下面是代码,其次是如何使用它的一个例子。

/// <summary>
/// Represents an instance of an IGrouping<>.  Used by GroupByMany(), GroupByWithRollup(), and GrandTotal().
/// </summary>
public class Grouping<TKey, TElement> : IGrouping<TKey, TElement>
{
    public TKey Key { get; set; }
    public IEnumerable<TElement> Items { get; set; }

    public IEnumerator<TElement> GetEnumerator()
    {
        return Items.GetEnumerator();
    }

    IEnumerator IEnumerable.GetEnumerator()
    {
        return Items.GetEnumerator();
    }
}

public static class Extensions
{
    /// <summary>
    /// Groups by two columns.
    /// </summary>
    /// <typeparam name="TElement">Type of elements to group.</typeparam>
    /// <typeparam name="TKey1">Type of the first expression to group by.</typeparam>
    /// <typeparam name="TKey2">Type of the second expression to group by.</typeparam>
    /// <param name="orderedElements">Elements to group.</param>
    /// <param name="groupByKey1Expression">The first expression to group by.</param>
    /// <param name="groupByKey2Expression">The second expression to group by.</param>
    /// <param name="newElementExpression">An expression that returns a new TElement.</param>
    public static IQueryable<Grouping<TKey1, TElement>> GroupByMany<TElement, TKey1, TKey2>(this IOrderedQueryable<TElement> orderedElements,
        Func<TElement, TKey1> groupByKey1Expression,
        Func<TElement, TKey2> groupByKey2Expression,
        Func<IGrouping<TKey1, TElement>, IGrouping<TKey2, TElement>, TElement> newElementExpression
        )
    {
        // Group the items by Key1 and Key2
        return from element in orderedElements
               group element by groupByKey1Expression(element) into groupByKey1
               select new Grouping<TKey1, TElement>
               {
                   Key = groupByKey1.Key,
                   Items = from key1Item in groupByKey1
                           group key1Item by groupByKey2Expression(key1Item) into groupByKey2
                           select newElementExpression(groupByKey1, groupByKey2)
               };
    }

    /// <summary>
    /// Returns a List of TElement containing all elements of orderedElements as well as subTotals and a grand total.
    /// </summary>
    /// <typeparam name="TElement">Type of elements to group.</typeparam>
    /// <typeparam name="TKey1">Type of the first expression to group by.</typeparam>
    /// <typeparam name="TKey2">Type of the second expression to group by.</typeparam>
    /// <param name="orderedElements">Elements to group.</param>
    /// <param name="groupByKey1Expression">The first expression to group by.</param>
    /// <param name="groupByKey2Expression">The second expression to group by.</param>
    /// <param name="newElementExpression">An expression that returns a new TElement.</param>
    /// <param name="subTotalExpression">An expression that returns a new TElement that represents a subTotal.</param>
    /// <param name="totalExpression">An expression that returns a new TElement that represents a grand total.</param>
    public static List<TElement> GroupByWithRollup<TElement, TKey1, TKey2>(this IOrderedQueryable<TElement> orderedElements,
        Func<TElement, TKey1> groupByKey1Expression,
        Func<TElement, TKey2> groupByKey2Expression,
        Func<IGrouping<TKey1, TElement>, IGrouping<TKey2, TElement>, TElement> newElementExpression,
        Func<IGrouping<TKey1, TElement>, TElement> subTotalExpression,
        Func<IQueryable<Grouping<TKey1, TElement>>, TElement> totalExpression
        )
    {
        // Group the items by Key1 and Key2
        IQueryable<Grouping<TKey1, TElement>> groupedItems = orderedElements.GroupByMany(groupByKey1Expression, groupByKey2Expression, newElementExpression);

        // Create a new list the items, subtotals, and the grand total.
        List<TElement> results = new List<TElement>();
        foreach (Grouping<TKey1, TElement> item in groupedItems)
        {
            // Add items under current group
            results.AddRange(item);
            // Add subTotal for current group
            results.Add(subTotalExpression(item));
        }
        // Add grand total
        results.Add(totalExpression(groupedItems));

        return results;
    }

    /// <summary>
    /// Returns the subTotal sum of sumExpression.
    /// </summary>
    /// <param name="sumExpression">An expression that returns the value to sum.</param>
    public static int SubTotal<TKey, TElement>(this IGrouping<TKey, TElement> query, Func<TElement, int> sumExpression)
    {
        return query.Sum(group => sumExpression(group));
    }

    /// <summary>
    /// Returns the subTotal sum of sumExpression.
    /// </summary>
    /// <param name="sumExpression">An expression that returns the value to sum.</param>
    public static decimal SubTotal<TKey, TElement>(this IGrouping<TKey, TElement> query, Func<TElement, decimal> sumExpression)
    {
        return query.Sum(group => sumExpression(group));
    }

    /// <summary>
    /// Returns the grand total sum of sumExpression.
    /// </summary>
    /// <param name="sumExpression">An expression that returns the value to sum.</param>
    public static int GrandTotal<TKey, TElement>(this IQueryable<Grouping<TKey, TElement>> query, Func<TElement, int> sumExpression)
    {
        return query.Sum(group => group.Sum(innerGroup => sumExpression(innerGroup)));
    }

    /// <summary>
    /// Returns the grand total sum of sumExpression.
    /// </summary>
    /// <param name="sumExpression">An expression that returns the value to sum.</param>
    public static decimal GrandTotal<TKey, TElement>(this IQueryable<Grouping<TKey, TElement>> query, Func<TElement, decimal> sumExpression)
    {
        return query.Sum(group => group.Sum(innerGroup => sumExpression(innerGroup)));
    }

并利用它的一个例子:

class Program
{
    static void Main(string[] args)
    {
        IQueryable<CustomObject> dataItems = (new[]
        {
            new CustomObject { City = "Seattle", Plan = "Plan B", Charges = 20 },
            new CustomObject { City = "Seattle", Plan = "Plan A", Charges = 10 },
            new CustomObject { City = "Seattle", Plan = "Plan B", Charges = 20 },
            new CustomObject { City = "Seattle", Plan = "Plan A", Charges = 10 },
            new CustomObject { City = "Seattle", Plan = "Plan A", Charges = 10 },
            new CustomObject { City = "Seattle", Plan = "Plan A", Charges = 10 },
            new CustomObject { City = "Portland", Plan = "Plan A", Charges = 10 },
            new CustomObject { City = "Portland", Plan = "Plan A", Charges = 10 },
            new CustomObject { City = "Portland", Plan = "Plan C", Charges = 30 },
            new CustomObject { City = "Portland", Plan = "Plan C", Charges = 30 },
            new CustomObject { City = "Portland", Plan = "Plan C", Charges = 30 }
        }).AsQueryable();

        List<CustomObject> results = dataItems.OrderBy(item => item.City).ThenBy(item => item.Plan).GroupByWithRollup(
            item => item.City,
            item => item.Plan,
            (primaryGrouping, secondaryGrouping) => new CustomObject
            {
                City = primaryGrouping.Key,
                Plan = secondaryGrouping.Key,
                Count = secondaryGrouping.Count(),
                Charges = secondaryGrouping.Sum(item => item.Charges)
            },
            item => new CustomObject
            {
                City = item.Key,
                Plan = "All",
                Count = item.SubTotal(subItem => subItem.Count),
                Charges = item.SubTotal(subItem => subItem.Charges)
            },
            items => new CustomObject
            {
                City = "All",
                Plan = "All",
                Count = items.GrandTotal(subItem => subItem.Count),
                Charges = items.GrandTotal(subItem => subItem.Charges)
            }
            );
        foreach (var result in results)
            Console.WriteLine(result);

        Console.Read();
    }
}

class CustomObject
{
    public string City { get; set; }
    public string Plan { get; set; }
    public int Count { get; set; }
    public decimal Charges { get; set; }

    public override string ToString()
    {
        return String.Format("{0} - {1} ({2} - {3})", City, Plan, Count, Charges);
    }
}


Answer 3:

@Ecyrb,你好从五年后!

我只是很模糊地熟悉LINQ to SQL的,超出标准的LINQ(对象)。 但是,因为你有一个“LINQ”标签从“LINQ-2-SQL”标签分开的,因为你似乎是在结果主要感兴趣的(而不是注册与数据库的变化),因为这是唯一想出了,当我在寻找一个LINQ相当于SQL Server的“汇总”分组功能的一派真正相关的资源,我会为当今任何人同样需要我自己的替代解决方案。

基本上,我的方法是创建一个 “.GroupBy()。ThenBy()” 可链接的语法类似 “.OrderBy()。ThenBy()” 语法。 我希望延长IGrouping对象的集合 - 你从运行得到“.GroupBy()”的结果 - 作为其源。 然后,它需要收集和取消组合他们回来在原来的对象分组之前。 最后,重新组的数据按照新的分组功能,制造另一组IGrouping的对象,并增加了新分组的对象到组源对象。

public static class mySampleExtensions {

    public static IEnumerable<IGrouping<TKey, TSource>> ThenBy<TSource, TKey> (     
        this IEnumerable<IGrouping<TKey, TSource>> source,
        Func<TSource, TKey> keySelector) {

        var unGroup = source.SelectMany(sm=> sm).Distinct(); // thank you flq at http://stackoverflow.com/questions/462879/convert-listlistt-into-listt-in-c-sharp
        var reGroup = unGroup.GroupBy(keySelector);

        return source.Concat(reGroup);}

}

您可以使用该方法通过把常量的值到“.ThenBy()”函数的相应区域,以匹配SQL服务器的汇总逻辑。 我更喜欢使用空值,因为它是铸造最灵活的恒定。 因为你在这两个.GroupBy()和.ThenBy(使用)的功能必须得到相同的对象类型转换是非常重要的。 使用您在08月31 '09你的第一反应创造了“dataItems”变量,它应该是这样的:

var rollItUp = dataItems
    .GroupBy(g=> new {g.City, g.Plan})
        .ThenBy(g=> new {g.City, Plan = (string) null})
        .ThenBy(g=> new {City = (string) null, Plan = (string) null})
    .Select(s=> new CustomObject {
        City = s.Key.City, 
        Plan = s.Key.Plan, 
        Count = s.Count(),
        Charges = s.Sum(a=> a.Charges)})
    .OrderBy(o=> o.City) // This line optional
        .ThenBy(o=> o.Plan); // This line optional

你可以用‘所有’替换“.ThenBy()”逻辑中的空白,你的愿望。

你可能会效仿SQL Server的分组集,也许立方体,用“).ThenBy(”的帮助。 此外,“.ThenBy()”是的工作对我很好,我不预见这个名字等同于任何问题“.ThenBy()”,“.OrderBy()”方法,因为它们有不同的签名,但如果有麻烦,你可能要考虑将其命名为“.ThenGroupBy()”来区分。

如前所述,我不使用LINQ到SQL,但我使用F#的类型提供系统,我的理解使用LINQ到SQL在许多方面引擎盖下。 所以,我想我这样的对象可拓从我的F#项目,它的工作原理如我所料。 虽然我完全不知道,如果这意味着什么有趣或者在这方面。



Answer 4:

颇为有趣的解决方案在这里提供

https://blogs.msdn.microsoft.com/mitsu/2007/12/21/playing-with-linq-grouping-groupbymany/

它描述了如何通过几个属性执行groupbby。 即:

var result = customers.GroupByMany(c => c.Country, c => c.City);

作为结果,你会得到可以简单地转换为平面列表层次结构。



文章来源: LINQ to SQL version of GROUP BY WITH ROLLUP