現(xiàn)有10W條數(shù)據(jù)，Access數(shù)據(jù)庫保存

通過正常提?。?br>

<% Set conn= Server.CreateObject("ADODB.Connection") connstr = "Provider=Microsoft.Jet.OLEDB.4.0;Data Source="&Server.MapPath("db2.mdb") conn.Open connstr Set rs = Server.CreateObject ("ADODB.Recordset") sql = "Select * from people order by id desc" rs.Open sql,conn,1,1 Do While Not rs.EOF     Response.write rs("id")&" | "     rs.MoveNext Loop %>

http://www.cnbruce.com/test/getrows/show1.asp

耗時3,250.000毫秒，總測試平均值在3秒左右

==========================================================

使用存儲過程提?。?br>

<% Set conn = Server.CreateObject("ADODB.Connection") Set cmd = Server.CreateObject("ADODB.Command") conn.Open "Provider=Microsoft.Jet.OLEDB.4.0; Data Source=" & Server.MapPath("db2.mdb") cmd.ActiveConnection = conn cmd.CommandText = "Select * from people order by id desc" Set rs = cmd.Execute Do While Not rs.EOF Response.write rs("id")&" | " rs.MoveNext Loop %>

http://www.cnbruce.com/test/getrows/show2.asp

耗時2,187.500毫秒，總測試平均值在2秒左右

=========================================================

以上兩種均不能徹底解決執(zhí)行時間漫長的問題，主要原因即是循環(huán)每次都須向數(shù)據(jù)庫抽取記錄（Command速度相對較快）

那么使用GetRows()方法呢：

<% Set conn = Server.CreateObject("ADODB.Connection") Set cmd = Server.CreateObject("ADODB.Command") conn.Open "Provider=Microsoft.Jet.OLEDB.4.0; Data Source=" & Server.MapPath("db2.mdb") cmd.ActiveConnection = conn cmd.CommandText = "Select * from people order by id desc" Set rs = cmd.Execute rsArray = rs.GetRows() For i = 0 To UBound(rsArray, 2) Response.Write rsArray(0, i)&" | " Next %>

http://www.cnbruce.com/test/getrows/show3.asp

耗時187.500毫秒，總測試平均值在0.2秒左右

GetRows()方法是將數(shù)據(jù)從 Recordset 復(fù)制到二維數(shù)組中，這是一個二維數(shù)組，第一個下標(biāo)標(biāo)識字段，第二個則標(biāo)識記錄號

所以rsArray = rs.GetRows()
rsArray(0, 0)就表示記錄集第一行的第一字段值
rsArray(1, 0)就表示記錄集第一行的第二字段值

數(shù)組的數(shù)據(jù)是保存在內(nèi)存中的，這就從根本上解決了每次顯示記錄還需向數(shù)據(jù)庫請求的麻煩。

另外，關(guān)于GetRows()的詳細(xì)介紹和用法，這里有份內(nèi)容參考

http://www.51windows.net/pages/Ado/mdmthgetrows.htm

http://www.blueidea.com/tech/program/2005/2853.asp