sql – Access 2007改进查询/ vba按链表分组

2019年8月4日 141次阅读

我收到一些表,其中有按链表分组的元素,我很难处理它.

该函数可以找到它,但我常常被问到从任务调度程序启动时它的宏在哪里或有一些内存问题.

我使用下面的代码找出idGroup(翻译成英文),我想知道是否有办法改进它,特别是它的速度,因为它需要长达一个小时的30 000行和大约2500组…(这就是我用VBA看进展的原因……)

'Simple example
'idGroup,id2,id1
'6338546,14322882,13608969
'6338546,13608969,13255363
'6338546,6338546,14322882
'6338546,11837926,11316332
'6338546,12297571,11837926
'6338546,13255363,12811071
'6338546,12811071,12297571
'6338546,7610194,7343817
'6338546,7935943,7610194
'6338546,8531387,7935943
'6338546,6944491,6611041
'6338546,7343817,6944491
'6338546,9968746,9632204
'6338546,10381694,9968746
'6338546,6611041,0
'6338546,8920224,8531387
'6338546,9632204,8920224
'6338546,11316332,10941093
'6338546,10941093,10381694


Public Function GetidGroup()
    'first id1 is always 0
    sql = "SELECT idGroup, id2, id1 FROM TABLE_WITH_LINKED_LIST WHERE id1='0' ORDER BY id2 DESC"
    Dim rs As Recordset
    Dim uidLikedList As String, id2 As String, id1 As String

    Set rs = CurrentDb.OpenRecordset(sql)
    Dim total As Long
    Dim idGroup As String
    Dim incrément As Long, progress As Double

    total = rs.RecordCount
    incrément = 1

    While Not rs.EOF
        progress = Math.Round(100 * incrément / total, 2)

        'Print in order to avoir freezing
        Debug.Print progress

        If rs.Fields("idGroup") = "" Then
            id2 = rs.Fields("id2")

            idGroup = precedentUid(id2)

            rs.Edit
            rs.Fields("idGroup") = idGroup
            rs.Update
        End If

        incrément = incrément + 1
        rs.MoveNext
    Wend

    rs.Close
    Set rs = Nothing
    GetidGroup = total
End Function

'Recursive function
'Deepest so far is about 62 calls
Public Function precedentUid(id2 As String) As String
    sql = "SELECT idGroup, id2 FROM TABLE_WITH_LINKED_LIST WHERE id1 = '" & id2 & "'"
    Dim rs As Recordset
    Dim precedentid2 As String
    Dim idGroup As String
    Dim ret As String

    Set rs = CurrentDb.OpenRecordset(sql)
    If rs.EOF Then
        rs.Close
        Set rs = Nothing
        precedentUid = id2
    Else
        'Some records have several references
        '56 impacted records  :
        'TODO : Give the min id2 to the group
        ret = "-1"
        While Not rs.EOF           
            If rs.Fields("idGroup") = "" Then
                precedentid2 = rs.Fields("id2")
                idGroup = precedentUid(precedentid2)

                If ret = "-1" Or CLng(ret) > CLng(idGroup) Then
                    ret = idGroup
                End If

                'Debug.Print id2 & " " & precedentid2 & " " & idGroup

                rs.Edit
                    rs.Fields("idGroup") = idGroup
                rs.Update
            End If
            rs.MoveNext
        Wend
        rs.Close
        Set rs = Nothing
        precedentUid = ret
    End If
End Function

最佳答案 一些建议：

>您正在打开大量记录集(对于precedentUid的每次调用).相反,请考虑使用按idGroup id1排序的单个记录集,并向上或向下搜索适当的值.
>由于您总是在idGroup id1上搜索,我建议将其作为主键.然后,您将能够使用Seek方法进行更快速的搜索.
>一旦拥有主键,就不需要编辑单个记录集,并且加载速度更快.必须更新idGroup时,请将SQL语句与CurrentDb.Execute一起使用.
>在字典中缓存搜索idGroup的结果(参考Microsoft Scripting Runtime in Tools – > References).这样,您不会在递归时重复搜索.
>您的示例数据似乎是所有数字,但您从记录集中将它们作为字符串进行检索.基础数据类型应为Long,而不是Text.如果您无法控制此情况,我会考虑使用适当的数据类型创建临时表.

点赞