VBA - 替换CSV中的逗号不在引号内

Question

  Filename = Dir(Filepath & "\" & "*.csv")
    While Filename <> ""
        SourceFile = Filepath & "\" & Filename
        TargetFile = SavePath & "\" & Replace(Filename, ".csv", ".txt")

        OpenAsUnicode = False

        Dim objFSO: Set objFSO = CreateObject("Scripting.FileSystemObject")
        'Detect Unicode Files
        Dim Stream: Set Stream = objFSO.OpenTextFile(SourceFile, 1, False)
        intChar1 = Asc(Stream.Read(1))
        intChar2 = Asc(Stream.Read(1))
        Stream.Close
        If intChar1 = 255 And intChar2 = 254 Then
            OpenAsUnicode = True
        End If

        'Get script content
        Set Stream = objFSO.OpenTextFile(SourceFile, 1, 0, OpenAsUnicode)
        arrData = Stream.ReadAll()
        Stream.Close

        'Create output file
        Dim objOut: Set objOut = objFSO.CreateTextFile(TargetFile)
        objOut.Write Replace(Replace(arrData,",", "#|#"), Chr(34), "")  '-- This line is working fine but it is replacing all the commas inside the text qualifier as well..
        objOut.Close

        Filename = Dir

    Wend

在上面的代码中，objOut.Write Replace(Replace(arrData,",", "#|#"), Chr(34), "") 行将所有逗号替换为#|#，包括里面的逗号string.so 我只想替换不在双引号中的逗号。

包含字符串的文件 "A","B,C",D

我需要的结果是 A#|#B,C#|#D

提前感谢您的帮助。

Answer 1

怎么样？

objOut.Write Mid(Replace(Replace(arrData,""",""", "#|#"), Chr(34), ""), 2)

基本上，现在将"," 交换为#|#。但这还不够，因为文件以" 开头。因此，正在使用Mid() 函数消除这个问题。如果文件也以" 结尾，那么您也必须对其进行调整。

基于 cmets 中提到的速度问题，这里是我用来测试此解决方案的完整代码：

Option Explicit
Option Compare Text

Public Sub ConvertFile()
Dim lngRowNumber As Long
Dim strLineFromFile As String
Dim strSourceFile As String
Dim strDestinationFile As String

strSourceFile = "C:\tmp\Extract.txt"
strDestinationFile = "C:\tmp\Extract_b.txt"

Open strSourceFile For Input As #1
Open strDestinationFile For Output As #2

lngRowNumber = 0

Do Until EOF(1)
    Line Input #1, strLineFromFile
    strLineFromFile = Mid(Replace(strLineFromFile, """,""", "#|#"), 2)
    Write #2, strLineFromFile
    strLineFromFile = vbNullString
Loop

Close #1
Close #2

End Sub

测试文件大小为 350 MB，行数略高于 400 万行。代码不到一分钟就完成了。