Question

我在SQL Server中有16000行，我想在物理上为记录分配文档编号并使用以下代码，当我的记录为8000时它工作正常但很懒但当记录增加到16000时它会给我一个超时错误，请帮我提高查询性能???

declare @NewDoc int;
set @NewDoc=0
declare t1_cursor cursor dynamic for select  documentheaderid,documentnumber from Accounting.DocumentHeader where FinancialPeriodFK=@FinancialPeriodFK and Date>=@FromDate and Date<=@ToDate order by Date
open t1_cursor

fetch next from t1_cursor
while(@@fetch_status=0)
begin
set @NewDoc=@NewDoc+1;


update  Accounting.DocumentHeader set DocumentNumber=@NewDoc where current of t1_cursor
fetch next from t1_cursor
end 
close t1_cursor
deallocate t1_cursor

Answer 1

尝试类似这样的事情 - 避免任何逐行激动行处理，如游标或while循环：

-- creates a temporary "inline" view of the data you're interested in and 
-- assigns each row a consecutive number
;WITH DataToUpdate AS
(
    SELECT
        DocumentHeaderId,
        DocumentNumber,
        NewDocNum = ROW_NUMBER() OVER(ORDER BY [Date])
    FROM
        Accounting.DocumentHeader
    WHERE
        FinancialPeriodFK = @FinancialPeriodFK
        AND Date >= @FromDate
        AND Date <= @ToDate
)
-- update the base table to make use of that new document number 
UPDATE dh
SET dh.DocumentNumber = dtu.NewDocNum
FROM Accounting.DocumentHeader dh
INNER JOIN DataToUpdate dtu ON dh.DocumentHeaderId = dtu.DocumentHeaderId

应该显着加快处理时间！

低速SQL查询

1 个答案: