Question

我正在尝试从常规查询创建rowcount-query，但生成的SQL似乎缺少GROUP BY导致错误的计数。有谁知道我做错了什么。

首先是查询：

var query = Session.QueryOver<InkoopFactuurListItem>()
    .Where(i => i.KlantId == Klant.Id)
    .AndRestrictionOn(i => i.Status).IsIn(statussen)
    .SelectList(list => list
        .SelectGroup(h => h.Id).WithAlias(() => dto.Id)
        .SelectGroup(h => h.Banknummer).WithAlias(() => dto.Banknummer)
        .SelectGroup(h => h.CrediteurNaam).WithAlias(() => dto.CrediteurNaam)
        .SelectGroup(h => h.DienstType).WithAlias(() => dto.DienstType)
        .SelectGroup(h => h.DocumentId).WithAlias(() => dto.DocumentId)
        .SelectGroup(h => h.DocumentNaam).WithAlias(() => dto.DocumentNaam)
        .SelectGroup(h => h.Factuurbedrag).WithAlias(() => dto.Factuurbedrag)
        .SelectGroup(h => h.Klantnummer).WithAlias(() => dto.Klantnummer)
        .SelectGroup(h => h.Factuurbtw).WithAlias(() => dto.Factuurbtw)
        .SelectGroup(h => h.FactuurDatum).WithAlias(() => dto.FactuurDatum)
        .SelectGroup(h => h.Factuurnummer).WithAlias(() => dto.Factuurnummer)
        .SelectGroup(h => h.IMSNummer).WithAlias(() => dto.IMSNummer)
        .SelectGroup(h => h.KlantId).WithAlias(() => dto.KlantId)
        .SelectGroup(h => h.Soortfactuur).WithAlias(() => dto.Soortfactuur)
        .SelectGroup(h => h.Status).WithAlias(() => dto.Status)
        .SelectGroup(h => h.VerwerktOp).WithAlias(() => dto.VerwerktOp)
        .SelectMin(h => h.Van).WithAlias(() => dto.Van)
        .SelectMax(h => h.Tot).WithAlias(() => dto.Tot))
    .TransformUsing(Transformers.AliasToBean<InkoopFactuurListItem>());

var rowcount = query.ToRowCountQuery().FutureValue<int>();

IEnumerable<InkoopFactuurListItem> results;
if (command.Page > 0 && command.PageSize > 0)
{
    results = query
        .Skip((command.Page - 1) * command.PageSize)
        .Take(command.PageSize).Future<InkoopFactuurListItem>();
}
else
{
    results = query
        .Take(command.PageSize)
        .Future<InkoopFactuurListItem>();
}

count = rowcount.Value;

生成的SQL：

SELECT count(*) as y0_ 
FROM vwInkoopFactuurListItem this_ 
WHERE this_.KlantId = @p0 and this_.Status in (@p1, @p2, @p3, @p4, @p5);

SELECT TOP (@p6) y0_, y1_, y2_, y3_, y4_, y5_, y6_, y7_, y8_, y9_, y10_, y11_, y12_, y13_, y14_, y15_, y16_, y17_ 
FROM (
    SELECT this_.InkoopFactuurId as y0_, this_.Banknummer as y1_, this_.CrediteurNaam as y2_, this_.DienstType as y3_, this_.DocumentId as y4_, this_.DocumentNaam as y5_, this_.Factuurbedrag as y6_, this_.Klantnummer as y7_, this_.Factuurbtw as y8_, this_.FactuurDatum as y9_, this_.Factuurnummer as y10_, this_.IMSNummer as y11_, this_.KlantId as y12_, this_.Soortfactuur as y13_, this_.Status as y14_, this_.VerwerktOp as y15_, min(this_.Van) as y16_, max(this_.Tot) as y17_,
    ROW_NUMBER() OVER(ORDER BY CURRENT_TIMESTAMP) as __hibernate_sort_row
    FROM vwInkoopFactuurListItem this_
    WHERE this_.KlantId = @p8 and this_.Status in (@p9, @p10, @p11, @p12, @p13)
    GROUP BY this_.InkoopFactuurId, this_.Banknummer, this_.CrediteurNaam, this_.DienstType, this_.DocumentId, this_.DocumentNaam, this_.Factuurbedrag, this_.Klantnummer, this_.Factuurbtw, this_.FactuurDatum, this_.Factuurnummer, this_.IMSNummer, this_.KlantId, this_.Soortfactuur, this_.Status, this_.VerwerktOp
) as query 
WHERE query.__hibernate_sort_row > @p7
ORDER BY query.__hibernate_sort_row; ;@p0 = 1 [Type: Int64 (0)], @p1 = 'OverigMatchingInkoop' [Type: String (255)], @p2 = 'OverigMatchingVerkoop' [Type: String (255)], @p3 = 'Overig' [Type: String (255)], @p4 = 'Geboekt' [Type: String (255)], @p5 = 'Geexporteerd' [Type: String (255)], @p6 = 10 [Type: Int32 (0)], @p7 = 284590 [Type: Int32 (0)], @p8 = 1 [Type: Int64 (0)], @p9 = 'OverigMatchingInkoop' [Type: String (255)], @p10 = 'OverigMatchingVerkoop' [Type: String (255)], @p11 = 'Overig' [Type: String (255)], @p12 = 'Geboekt' [Type: String (255)], @p13 = 'Geexporteerd' [Type: String (255)]

那么rowcount-query中的GROUP BY在哪里？

更新事实证明，ToRowCountQuery从原始查询中删除选择和分组。那我怎么能这样做：

select count(*)
from (... query ...)

Firo的答案似乎是朝着正确方向迈出的一步，但我无法使用CountDistinct，因为我需要对所有字段进行计数，包括选择列表中显示的最小/最大字段。接下来，CountDistinct不会将IProjection作为参数：

//
// Summary:
//     A distinct property value count
public static CountProjection CountDistinct(Expression<Func<object>> expression);
//
// Summary:
//     A distinct property value count
public static CountProjection CountDistinct<T>(Expression<Func<T, object>> expression);
//
// Summary:
//     A distinct property value count
//
// Parameters:
//   propertyName:
public static CountProjection CountDistinct(string propertyName);

更新2

根据Firo答案中的链接，我提出了这个替代方案。我将query修改为分离的QueryOver对象。接下来，我尝试了这个：

var rowcount = Session.QueryOver<InkoopFactuurListItem>()
    .Select(
        Projections.Alias(
            Projections.Count(Projections.SubQuery(detachedQuery)), "count"
        )
    )
    .FutureValue<int>();

但是，这会产生ArgumentOutOfRangeException：

Index was out of range. Must be non-negative and less than the size of the collection.
Parameter name: index

Answer 1

编辑：

ToRowCountQuery即使是分组也会删除所有预测。如果您自己运行查询，则会在count(*)的查询中看到group by将返回每个组的计数而不是组的数量。您必须COUNT DISTINCT所有组列才能获得组数。

你需要像SELECT count(*) FROM (query)这样的东西，这是不可能的。我只能想到一个脆弱的解决方案，你可以在其中生成如here所示的sql然后

string sql = GenerateSQL(query.UnderlyingCriteria);
rowcount = session.CreateSQLQuery(string.Format("SELECT Count(*) FROM ({0})", sql)).SetParameter("???", KlantId).FutureValue<int>();

ToRowCountQuery似乎忽略了分组

1 个答案: