在 SQL Server 中分页时保留总结果信息

Question

我正在尝试在 SQL Server 的存储过程中进行分页，如下所示：

/*Assign a row number to each row*/
SELECT
ROW_NUMBER() OVER (ORDER BY A, B, C ASC) AS ROW_NUMBER,
A, B, C
FROM ABC
WHERE ROW_NUMBER BETWEEN @startRecordNumber and @endRecordNumber

但是，我的调用代码还想知道分页前的原始结果集中有多少结果。所以，我把我的代码改成这样：

/*Put the results into a temp table first*/
SELECT
ROW_NUMBER() OVER (ORDER BY A, B, C ASC) AS ROW_NUMBER,
A, B, C
INTO #TOTAL_RESULTS
FROM ABC

/*Get the total results from @@RowCount*/
declare @totalResults bigint
set @totalResults = @@RowCount

/*Now just get the desired page from the temp table*/
SELECT
A, B, C
FROM #TOTAL_RESULTS
WHERE ROW_NUMBER BETWEEN @startRecordNumber and @endRecordNumber

DROP TABLE #TOTAL_RESULTS

这对我来说感觉很迂回。有什么方法可以在不制作临时表的情况下获得原始结果集的大小？也许只是一个公用表表达式？我似乎想不出办法。

如果重要的话，这里是 ABC 的架构：

ABC
A（PK，smallint 不为空）
B（PK，smallint 不为空）
C (PK, smallint 不为空)

Answer 1

您可以使用窗口聚合函数。即COUNT(*) OVER() 下面的示例。

;WITH cte As
(
SELECT *, 
        ROW_NUMBER() OVER (ORDER BY number) AS RN,
        COUNT(*) OVER() AS Cnt
FROM master..spt_values
)
SELECT *
FROM cte 
WHERE RN BETWEEN 101 and 200

或者- 不是一个严肃的建议，但确实避免了线轴、工作台和双重排序:-)

DECLARE @Spid INT = @@Spid
DECLARE @TraceID INT

DECLARE @maxfilesize BIGINT = 5
DECLARE @filepath NVARCHAR(200) = N'C:\trace_' + LEFT(NEWID(),36)

EXEC sp_trace_create @TraceID OUTPUT, 0, @filepath, @maxfilesize, NULL 

exec sp_trace_setevent @TraceID, 146, 1, 1
exec sp_trace_setevent @TraceID, 146, 22, 1
exec sp_trace_setevent @TraceID, 146, 34, 1
exec sp_trace_setevent @TraceID, 146, 51, 1
exec sp_trace_setevent @TraceID, 146, 12, 1
-- filter for spid
EXEC sp_trace_setfilter @TraceID, 12, 0, 0, @Spid
-- start the trace
EXEC sp_trace_setstatus @TraceID, 1


;WITH cte AS
(
SELECT number, type, name, 
        ROW_NUMBER() OVER (ORDER BY number, type, name) AS RN
FROM master..spt_values
)
SELECT * FROM cte
WHERE RN BETWEEN 101 AND 200
OR RN+1 =0 /*To stop a "TOP 200" getting added to the plan*/

;WITH  XMLNAMESPACES ('http://schemas.microsoft.com/sqlserver/2004/07/showplan' as sql)
     SELECT  ActualRows
         FROM   fn_trace_getinfo(@TraceID) fn
                CROSS APPLY fn_trace_gettable(CAST(value AS NVARCHAR(200)), 1)
                CROSS APPLY (SELECT CAST(TextData AS XML) AS xPlan) x
                CROSS APPLY (SELECT T.relop.value('@ActualRows', 'INT') AS ActualRows
                             FROM   xPlan.nodes('//sql:RelOp[@LogicalOp="Segment"]/sql:RunTimeInformation/sql:RunTimeCountersPerThread') T(relop)) ca
         WHERE  property = 2
                AND ObjectName<>'fn_trace_getinfo' AND TextData NOT LIKE '%ThisQuery%'  

-- Stop the trace
EXEC sp_trace_setstatus @TraceID, 0
-- Close and delete the trace
EXEC sp_trace_setstatus @TraceID, 2

Answer 2

据我所知，SQL Server 不可能同时返回行和分配变量。因此，如果您希望将行数存储到变量中，则必须以一种或另一种方式拥有多个语句。

但如果可以将总计作为列返回，并且表中的 A、B 和 C 列没有重复项，则可以返回总计，例如，如下所示：

SELECT
  A, B, C,
  TotalResults = RowNumAsc + RowNumDesc - 1
FROM (
  SELECT
    A, B, C,
    RowNumAsc = ROW_NUMBER() OVER (ORDER BY A, B, C),
    RowNumDesc = ROW_NUMBER() OVER (ORDER BY A DESC, B DESC, C DESC)
  FROM
) s
WHERE RowNumAsc BETWEEN @startRecordNumber AND @endRecordNumber