Question

这里有一个代码2 DataTables。它们都填充了SQL Server 2012中的SqlDataAdabter。dtParticipants中有15.000行，dtSubjectResult中有80.000行。

foreach (DataRow rowPart in dtParticipants.Rows)
{
    int partId = int.Parse(rowPart["main_id"].ToString());
    string partUserCode = rowPart["o_user_code"].ToString();

    DataRow[] filteredRows = dtSubjectResult.Select("main_id = " + partId);
    if (filteredRows == null)
        continue;

    foreach (DataRow row in filteredRows)
    {
        row["correct_count"] = 5;//unit.CorrectCount;
        row["incorrect_count"] = 15;//unit.IncorrectCount;
        row["empty_count"] = 20;//unit.EmptyCount;
        row["net_count"] = 12;//unit.Total;
    }//foreach 2
}//foreach 1

但是这段代码非常慢......当我在第二个foreach中注释代码时，代码运行非常正常，没有任何速度问题。

有趣的是，当我尝试这样的代码时，它的效果非常好。

foreach (DataRow row in dtSubjectResult.Rows)
{
    row["correct_count"] = 5;//unit.CorrectCount;
    row["incorrect_count"] = 15;//unit.IncorrectCount;
    row["empty_count"] = 20;//unit.EmptyCount;
    row["net_count"] = 12;//unit.Total;
}

那么问题是什么... IS循环过滤的行慢？我不知道??

Answer 1

多重过滤操作是一个严重的性能损失

DataRow[] filteredRows = dtSubjectResult.Select("main_id = " + partId);

Select需要解释过滤器表达式，在每次15k迭代中扫描dtSubjectResult中的80k行

Join 2个表并一次更新dtSubjectResult行：

var grouped = dtParticipants.AsEnumerable()
    .Join(dtSubjectResult.AsEnumerable(),
          rPart => (int)rPart["main_id"],
          rSubj => (int)rSubj["main_id"],
          (rPart, rSubj) => new { P = rPart, S = rSubj })
    .GroupBy(x => x.P["main_id"]);

foreach (var gr in grouped)
{
    foreach (var row in gr)
    {
        row.S["correct_count"] = 5;//unit.CorrectCount;
        row.S["incorrect_count"] = 15;//unit.IncorrectCount;
        row.S["empty_count"] = 20;//unit.EmptyCount;
        row.S["net_count"] = 12;//unit.Total;
    }
}

在c＃中更改DataTable的过滤DataRows非常慢

1 个答案: