在c#中是否有默认/官方/推荐的方法来解析CSV文件?我不想滚动自己的解析器。

另外,我也见过人们使用ODBC/OLE DB通过文本驱动程序读取CSV的实例,很多人因为它的“缺点”而不鼓励这样做。这些缺点是什么?

理想情况下,我正在寻找一种方法,通过它我可以通过列名读取CSV,使用第一个记录作为报头/字段名。给出的一些答案是正确的,但基本上是将文件反序列化为类。


当前回答

让一个库为你处理所有的细节!: -)

检查FileHelpers和保持干燥-不重复自己-不需要重新发明轮子的亿万次....

基本上,您只需要定义数据的形状——CSV中各个行中的字段——通过一个公共类(以及诸如默认值、NULL值替换等经过精心考虑的属性),将FileHelpers引擎指向一个文件,然后就可以从该文件中获得所有条目。一个简单的操作-卓越的性能!

其他回答

让一个库为你处理所有的细节!: -)

检查FileHelpers和保持干燥-不重复自己-不需要重新发明轮子的亿万次....

基本上,您只需要定义数据的形状——CSV中各个行中的字段——通过一个公共类(以及诸如默认值、NULL值替换等经过精心考虑的属性),将FileHelpers引擎指向一个文件,然后就可以从该文件中获得所有条目。一个简单的操作-卓越的性能!

如果任何人想要一个代码片段,他们可以直接输入自己的代码,而不必绑定库或下载包。以下是我写的一个版本:

    public static string FormatCSV(List<string> parts)
    {
        string result = "";

        foreach (string s in parts)
        {
            if (result.Length > 0)
            {
                result += ",";

                if (s.Length == 0)
                    continue;
            }

            if (s.Length > 0)
            {
                result += "\"" + s.Replace("\"", "\"\"") + "\"";
            }
            else
            {
                // cannot output double quotes since its considered an escape for a quote
                result += ",";
            }
        }

        return result;
    }

    enum CSVMode
    {
        CLOSED = 0,
        OPENED_RAW = 1,
        OPENED_QUOTE = 2
    }

    public static List<string> ParseCSV(string input)
    {
        List<string> results;

        CSVMode mode;

        char[] letters;

        string content;


        mode = CSVMode.CLOSED;

        content = "";
        results = new List<string>();
        letters = input.ToCharArray();

        for (int i = 0; i < letters.Length; i++)
        {
            char letter = letters[i];
            char nextLetter = '\0';

            if (i < letters.Length - 1)
                nextLetter = letters[i + 1];

            // If its a quote character
            if (letter == '"')
            {
                // If that next letter is a quote
                if (nextLetter == '"' && mode == CSVMode.OPENED_QUOTE)
                {
                    // Then this quote is escaped and should be added to the content

                    content += letter;

                    // Skip the escape character
                    i++;
                    continue;
                }
                else
                {
                    // otherwise its not an escaped quote and is an opening or closing one
                    // Character is skipped

                    // If it was open, then close it
                    if (mode == CSVMode.OPENED_QUOTE)
                    {
                        results.Add(content);

                        // reset the content
                        content = "";

                        mode = CSVMode.CLOSED;

                        // If there is a next letter available
                        if (nextLetter != '\0')
                        {
                            // If it is a comma
                            if (nextLetter == ',')
                            {
                                i++;
                                continue;
                            }
                            else
                            {
                                throw new Exception("Expected comma. Found: " + nextLetter);
                            }
                        }
                    }
                    else if (mode == CSVMode.OPENED_RAW)
                    {
                        // If it was opened raw, then just add the quote 
                        content += letter;
                    }
                    else if (mode == CSVMode.CLOSED)
                    {
                        // Otherwise open it as a quote 

                        mode = CSVMode.OPENED_QUOTE;
                    }
                }
            }
            // If its a comma seperator
            else if (letter == ',')
            {
                // If in quote mode
                if (mode == CSVMode.OPENED_QUOTE)
                {
                    // Just read it
                    content += letter;
                }
                // If raw, then close the content
                else if (mode == CSVMode.OPENED_RAW)
                {
                    results.Add(content);

                    content = "";

                    mode = CSVMode.CLOSED;
                }
                // If it was closed, then open it raw
                else if (mode == CSVMode.CLOSED)
                {
                    mode = CSVMode.OPENED_RAW;

                    results.Add(content);

                    content = "";
                }
            }
            else
            {
                // If opened quote, just read it
                if (mode == CSVMode.OPENED_QUOTE)
                {
                    content += letter;
                }
                // If opened raw, then read it
                else if (mode == CSVMode.OPENED_RAW)
                {
                    content += letter;
                }
                // It closed, then open raw
                else if (mode == CSVMode.CLOSED)
                {
                    mode = CSVMode.OPENED_RAW;

                    content += letter;
                }
            }
        }

        // If it was still reading when the buffer finished
        if (mode != CSVMode.CLOSED)
        {
            results.Add(content);
        }

        return results;
    }

这里有一个简短而简单的解决方案。

                using (TextFieldParser parser = new TextFieldParser(outputLocation))
                 {
                        parser.TextFieldType = FieldType.Delimited;
                        parser.SetDelimiters(",");
                        string[] headers = parser.ReadLine().Split(',');
                        foreach (string header in headers)
                        {
                            dataTable.Columns.Add(header);
                        }
                        while (!parser.EndOfData)
                        {
                            string[] fields = parser.ReadFields();
                            dataTable.Rows.Add(fields);
                        }
                    }

如果你只需要读取csv文件,那么我推荐这个库:一个快速csv阅读器 如果你还需要生成csv文件,那么使用FileHelpers

它们都是免费和开源的。

这个解析器支持在列中嵌套逗号和引号:

static class CSVParser
{
    public static string[] ParseLine(string line)
    {
        List<string> cols = new List<string>();
        string value = null;

        for(int i = 0; i < line.Length; i++)
        {
            switch(line[i])
            {
                case ',':
                    cols.Add(value);
                    value = null;
                    if(i == line.Length - 1)
                    {// It ends with comma
                        cols.Add(null);
                    }
                    break;
                case '"':
                    cols.Add(ParseEnclosedColumn(line, ref i));
                    i++;
                    break;
                default:
                    value += line[i];
                    if (i == line.Length - 1)
                    {// Last character
                        cols.Add(value);                           
                    }
                    break;
            }
        }

        return cols.ToArray();
    }//ParseLine

    static string ParseEnclosedColumn(string line, ref int index)
    {// Example: "b"",bb"
        string value = null;
        int numberQuotes = 1;
        int index2 = index;

        for (int i = index + 1; i < line.Length; i++)
        {
            index2 = i;
            switch (line[i])
            {
                case '"':
                    numberQuotes++;
                    if (numberQuotes % 2 == 0)
                    {
                        if (i < line.Length - 1 && line[i + 1] == ',')
                        {
                            index = i;
                            return value;
                        }
                    }
                    else if (i > index + 1 && line[i - 1] == '"')
                    {
                        value += '"';
                    }
                    break;
                default:
                    value += line[i];
                    break;
            }
        }

        index = index2;
        return value;
    }//ParseEnclosedColumn 
}//class CSVParser