1

为了获得排序后的聚合字符串，我在下面编写了 CLR 函数。但是，它总是返回空而不是我预期的，就像“001,002,003”一样。我试图在visual studio 2017中调试CLR函数，但抛出了错误信息

操作无法完成。未指定的错误

代码：

[Serializable]
[SqlUserDefinedAggregate(
    Format.UserDefined, //use clr serialization to serialize the intermediate result
    Name = "CLRSortedCssvAgg", //aggregate name on sql
    IsInvariantToNulls = true, //optimizer property
    IsInvariantToDuplicates = false, //optimizer property
    IsInvariantToOrder = false, //optimizer property
    IsNullIfEmpty = false, //optimizer property
    MaxByteSize = -1) //maximum size in bytes of persisted value
]

public class SortedCssvConcatenateAgg : IBinarySerialize
{
    /// <summary>
    /// The variable that holds all the strings to be aggregated.
    /// </summary>
    List<string> aggregationList;

    StringBuilder accumulator;

    /// <summary>
    /// Separator between concatenated values.
    /// </summary>
    const string CommaSpaceSeparator = ", ";

    /// <summary>
    /// Initialize the internal data structures.
    /// </summary>
    public void Init()
    {
        accumulator = new StringBuilder();
        aggregationList = new List<string>();
    }

    /// <summary>
    /// Accumulate the next value, not if the value is null or empty.
    /// </summary>
    public void Accumulate(SqlString value)
    {
        if (value.IsNull || String.IsNullOrEmpty(value.Value))
        {
            return;
        }

        aggregationList.Add(value.Value);
    }

    /// <summary>
    /// Merge the partially computed aggregate with this aggregate.
    /// </summary>
    /// <param name="other"></param>
    public void Merge(SortedCssvConcatenateAgg other)
    {
        aggregationList.AddRange(other.aggregationList);
    }

    /// <summary>
    /// Called at the end of aggregation, to return the results of the aggregation.
    /// </summary>
    /// <returns></returns>
    public SqlString Terminate()
    {
        if (aggregationList != null && aggregationList.Count > 0)
        {
            aggregationList.Sort();
            accumulator.Append(string.Join(CommaSpaceSeparator, aggregationList));
            aggregationList.Clear();
        }

        return new SqlString(accumulator.ToString());
    }

    public void Read(BinaryReader r)
    {
        accumulator = new StringBuilder(r.ReadString());
    }

    public void Write(BinaryWriter w)
    {
        w.Write(accumulator.ToString());
    }
}
4

2 回答 2

2

你很亲密。只需要一些小的调整。执行以下操作，它将起作用（我对其进行了测试）：

  1. 删除所有对accumulator. 它不被使用。

  2. 将 Terminate()、Read() 和 Write() 方法替换为以下内容： 

    public SqlString Terminate()
{
    string _Aggregation = null;

    if (aggregationList != null && aggregationList.Count > 0)
    {
        aggregationList.Sort();
        _Aggregation = string.Join(CommaSpaceSeparator, aggregationList);
    }

    return new SqlString(_Aggregation);
}

public void Read(BinaryReader r)
{
    int _Count = r.ReadInt32();
    aggregationList = new List<string>(_Count);

    for (int _Index = 0; _Index < _Count; _Index++)
    {
        aggregationList.Add(r.ReadString());
    }
}

public void Write(BinaryWriter w)
{
    w.Write(aggregationList.Count);
    foreach (string _Item in aggregationList)
    {
        w.Write(_Item);
    }
}

也就是说，我不确定这种方法是否比这种方法更快或更慢FOR XML，但 UDA 肯定会产生更易读的查询，尤其是在您需要多个聚合的情况下。

不过，我应该提一下，从 SQL Server 2017 开始，这变成了一个内置函数：STRING_AGG（它允许通过WITHIN GROUP (ORDER BY ... )子句进行排序）。

于 2018-09-19T06:26:23.257 回答
0

In your Accumulate and Merge, you're dealing with your aggregationList; in Read and Write you're dealing with accumulator. You should pick one or the other for all of them and use it. As I understand it, Read and Write are used when the engine needs to persist temporary results to a work table. For your case, when it does that, it's persisting only your empty StringBuilder.

于 2018-05-03T14:58:34.717 回答