c# - 添加一种在 CSV 中将逗号保留到 DataTable 函数的方法

Question

我有一个将 .csv 文件转换为数据表的函数。我正在转换的列之一是其中有一个逗号的名称字段，即“Doe, John”在转换函数时将其视为 2 个单独的字段，因为逗号。我需要数据表将其保存为数据表中的一个字段 Doe、John。

Function CSV2DataTable(ByVal filename As String, ByVal sepChar As String) As DataTable
    Dim reader As System.IO.StreamReader
    Dim table As New DataTable
    Dim colAdded As Boolean = False

    Try
        ''# open a reader for the input file, and read line by line
        reader = New System.IO.StreamReader(filename)
        Do While reader.Peek() >= 0
            ''# read a line and split it into tokens, divided by the specified 
            ''# separators
            Dim tokens As String() = System.Text.RegularExpressions.Regex.Split _
                (reader.ReadLine(), sepChar)
            ''# add the columns if this is the first line
            If Not colAdded Then
                For Each token As String In tokens
                    table.Columns.Add(token)
                Next
                colAdded = True
            Else
                ''# create a new empty row
                Dim row As DataRow = table.NewRow()
                ''# fill the new row with the token extracted from the current 
                ''# line
                For i As Integer = 0 To table.Columns.Count - 1
                    row(i) = tokens(i)
                Next
                ''# add the row to the DataTable
                table.Rows.Add(row)
            End If
        Loop

        Return table
    Finally
        If Not reader Is Nothing Then reader.Close()
    End Try
End Function

Answer 1

不要使用 .Split() 函数来读取您的 csv 数据。它不仅会导致您刚刚遇到的那种错误，而且速度也会变慢。您需要一个基于状态机 的解析器。这样会更快，并且更容易正确处理引号括起的文本。

我这里有一个例子:
Reading CSV files in C#

您还可以使用 codeplex 上备受推崇的 CSV 阅读器:
http://www.codeproject.com/KB/database/CsvReader.aspx

---

你会像这样使用我的代码:

Function DataTableFromCSV(ByVal filename As String) As DataTable
    Dim table As New DataTable
    Dim colAdded As Boolean = False

    For Each record As IList(Of String) In CSV.FromFile(filename)
        ''# Add column headers on first iteration
        If Not colAdded Then
            For Each token As String In record
                table.Columns.Add(token)
            Next token
            colAdded = True
        Else
            ''# add the row to the table
            Dim row As DataRow = table.NewRow()
            For i As Integer = 0 To table.Columns.Count - 1
                row(i) = record(i)
            Next
            table.Rows.Add(row)
        End If
    Next record

    Return table
End Function  

如果您使用的是 .net 3.5 或更高版本，我会稍微不同地编写它以将列创建从 for each 循环中拉出来(使用类型推断和 .Take(1) )，但我想确保这也适用于 .Net 2.0。