我在尝试为需要按属性分组的数组开发解决方案时遇到困难,我认为这必须非常简单,但是...
我有一个必须在按状态分组的批处理过程中处理的供应商列表。
public class SupplierDetails
{
public int SupplierId { get; set; }
public string Status { get; set; }
}
供应商列表将具有不同的状态和 SupplierId,供应商需要按照它们插入列表的顺序进行处理,因此,我需要向流程发送一组具有相同状态的第一批供应商,一旦处理完毕,我必须将下一个具有相同状态的供应商发送到该流程,依此类推...
列表的样例
[SupplierId=1,Status='C'],[SupplierId=2,Status='C'],[SupplierId=3,Status='A'],[SupplierId=4,Status='C']
结果列表可以让我以这种方式将 Suppliers 发送到 BatchProcess
BatchProcess([SupplierId=1,Status='C'],[SupplierId=2,Status='C'])
BatchProcess([SupplierId=3,Status='A'])
BatchProcess([SupplierId=4,Status='C'])
另一种解释是......“所以当你有 AAABBBABABA 时,你会想要一组中的前三个,然后是一组中的下三个,然后是一堆单独的项目”感谢@Chris
现在,我开发的解决方案不是很漂亮,但使用了泛型
public List<IEnumerable<T>> FlattenBatchedArray<T, Y>(IEnumerable<T> arr, Func<T, Y> getPropertyValue)
{
if (arr == null || arr.Count() == 0 || getPropertyValue == null)
return null;
var listArr = arr.ToArray();
int indexSkip = 0;
int indexTake = 0;
var groupedList = new List<IEnumerable<T>>();
while (indexTake < listArr.Length)
{
indexSkip = indexTake;
var currentValue = getPropertyValue(listArr.ElementAt(indexSkip));
indexTake = Array.FindIndex(listArr, indexSkip, x => !getPropertyValue(x).Equals(currentValue));
if (indexTake == -1)
indexTake = listArr.Length;
var next = listArr.Skip(indexSkip).Take(indexTake - indexSkip);
groupedList.Add(next.ToList());
}
return groupedList;
}
这个函数通过了我需要的所有测试
[TestMethod]
public void TestMethod3()
{
var listPO = new List<SupplierDetails>();
listPO.Add(new UnitTestProject2.SupplierDetails() { SupplierId = 1, Status = "O" });
listPO.Add(new UnitTestProject2.SupplierDetails() { SupplierId = 2, Status = "C" });
listPO.Add(new UnitTestProject2.SupplierDetails() { SupplierId = 3, Status = "C" });
listPO.Add(new UnitTestProject2.SupplierDetails() { SupplierId = 4, Status = "O" });
listPO.Add(new UnitTestProject2.SupplierDetails() { SupplierId = 5, Status = "O" });
listPO.Add(new UnitTestProject2.SupplierDetails() { SupplierId = 6, Status = "O" });
listPO.Add(new UnitTestProject2.SupplierDetails() { SupplierId = 7, Status = "C" });
listPO.Add(new UnitTestProject2.SupplierDetails() { SupplierId = 8, Status = "O" });
var result = FlattenBatchedArray(listPO, (x) => x.Status);
Assert.AreEqual(result.Count, 5);
Assert.AreEqual(result[0].Count(), 1);
Assert.AreEqual(result[1].Count(), 2);
Assert.AreEqual(result[2].Count(), 3);
Assert.AreEqual(result[3].Count(), 1);
Assert.AreEqual(result[4].Count(), 1);
Assert.AreEqual(result[0].ElementAt(0).SupplierId, 1);
Assert.AreEqual(result[0].ElementAt(0).Status, "O");
Assert.AreEqual(result[1].ElementAt(0).SupplierId, 2);
Assert.AreEqual(result[1].ElementAt(0).Status, "C");
Assert.AreEqual(result[1].ElementAt(1).SupplierId, 3);
Assert.AreEqual(result[1].ElementAt(1).Status, "C");
Assert.AreEqual(result[2].ElementAt(0).SupplierId, 4);
Assert.AreEqual(result[2].ElementAt(0).Status, "O");
Assert.AreEqual(result[2].ElementAt(1).SupplierId, 5);
Assert.AreEqual(result[2].ElementAt(1).Status, "O");
Assert.AreEqual(result[2].ElementAt(2).SupplierId, 6);
Assert.AreEqual(result[2].ElementAt(2).Status, "O");
Assert.AreEqual(result[3].ElementAt(0).SupplierId, 7);
Assert.AreEqual(result[3].ElementAt(0).Status, "C");
Assert.AreEqual(result[4].ElementAt(0).SupplierId, 8);
Assert.AreEqual(result[4].ElementAt(0).Status, "O");
listPO = new List<SupplierDetails>();
listPO.Add(new UnitTestProject2.SupplierDetails() { SupplierId = 1, Status = "O" });
result = FlattenBatchedArray(listPO, (x) => x.Status);
Assert.AreEqual(result.Count, 1);
listPO = new List<SupplierDetails>();
listPO.Add(new UnitTestProject2.SupplierDetails() { SupplierId = 1, Status = "O" });
listPO.Add(new UnitTestProject2.SupplierDetails() { SupplierId = 2, Status = "O" });
result = FlattenBatchedArray(listPO, (x) => x.Status);
Assert.AreEqual(result.Count, 1);
Assert.AreEqual(result[0].Count(), 2);
listPO = new List<SupplierDetails>();
listPO.Add(new UnitTestProject2.SupplierDetails() { SupplierId = 1, Status = "O" });
listPO.Add(new UnitTestProject2.SupplierDetails() { SupplierId = 2, Status = "C" });
result = FlattenBatchedArray(listPO, (x) => x.Status);
Assert.AreEqual(result.Count, 2);
Assert.AreEqual(result[0].Count(), 1);
Assert.AreEqual(result[1].Count(), 1);
Assert.AreEqual(result[0].ElementAt(0).SupplierId, 1);
Assert.AreEqual(result[0].ElementAt(0).Status, "O");
Assert.AreEqual(result[1].ElementAt(0).SupplierId, 2);
Assert.AreEqual(result[1].ElementAt(0).Status, "C");
listPO = new List<SupplierDetails>();
listPO.Add(new UnitTestProject2.SupplierDetails() { SupplierId = 1, Status = "O" });
listPO.Add(new UnitTestProject2.SupplierDetails() { SupplierId = 2, Status = "O" });
listPO.Add(new UnitTestProject2.SupplierDetails() { SupplierId = 3, Status = "C" });
result = FlattenBatchedArray(listPO, (x) => x.Status);
Assert.AreEqual(result.Count, 2);
Assert.AreEqual(result[0].Count(), 2);
Assert.AreEqual(result[1].Count(), 1);
Assert.AreEqual(result[0].ElementAt(0).SupplierId, 1);
Assert.AreEqual(result[0].ElementAt(0).Status, "O");
Assert.AreEqual(result[0].ElementAt(1).SupplierId, 2);
Assert.AreEqual(result[0].ElementAt(1).Status, "O");
Assert.AreEqual(result[1].ElementAt(0).SupplierId, 3);
Assert.AreEqual(result[1].ElementAt(0).Status, "C");
listPO = new List<SupplierDetails>();
listPO.Add(new UnitTestProject2.SupplierDetails() { SupplierId = 1, Status = "O" });
listPO.Add(new UnitTestProject2.SupplierDetails() { SupplierId = 2, Status = "C" });
listPO.Add(new UnitTestProject2.SupplierDetails() { SupplierId = 3, Status = "C" });
result = FlattenBatchedArray(listPO, (x) => x.Status);
Assert.AreEqual(result.Count, 2);
Assert.AreEqual(result[0].Count(), 1);
Assert.AreEqual(result[1].Count(), 2);
Assert.AreEqual(result[0].ElementAt(0).SupplierId, 1);
Assert.AreEqual(result[0].ElementAt(0).Status, "O");
Assert.AreEqual(result[1].ElementAt(0).SupplierId, 2);
Assert.AreEqual(result[1].ElementAt(0).Status, "C");
Assert.AreEqual(result[1].ElementAt(1).SupplierId, 3);
Assert.AreEqual(result[1].ElementAt(1).Status, "C");
}
最佳答案
这是一个不会多次枚举输入序列的解决方案:
public static IEnumerable<IEnumerable<TSource>> Batch<TKey, TSource>(IEnumerable<TSource> source, Func<TSource, TKey> keySelector)
{
var buffer = new List<TSource>();
foreach (var element in source)
{
if ((buffer.Count > 0) && !Equals(keySelector(element), keySelector(buffer.Last())))
{
yield return buffer.ToArray();
buffer.Clear();
}
buffer.Add(element);
}
if (buffer.Count > 0)
yield return buffer.ToArray();
}
(这也处理其中一个键为空的情况,尽管这可能不是必需的。)
[编辑:在意识到我不需要 prev
变量后我简化了它,因为我可以只使用 buffer.Last()
。]
关于c# - 批量分组数组,我们在Stack Overflow上找到一个类似的问题: https://stackoverflow.com/questions/46444680/