Question

除了引号之间的列之外，我需要用逗号分割csv文件。但是，我在这里所看到的似乎并没有实现我所需要的，而且列中的逗号被分成单独的数组项。

    public List<string> GetData(string dataFile, int row)
    {
        try
        {
            var lines = File.ReadAllLines(dataFile).Select(a => a.Split(';'));
            var csv = from line in lines select (from piece in line select piece.Split(',')).ToList();
            var foo = csv.ToList();
            var result = foo[row][0].ToList();
            return result;
        }
        catch
        {
            return null;
        }
    }

    private const string QUOTE = "\"";
    private const string ESCAPED_QUOTE = "\"\"";
    private static char[] CHARACTERS_THAT_MUST_BE_QUOTED = { ',', '"', '\n' };

    public static string Escape(string s)
    {
        if (s.Contains(QUOTE))
            s = s.Replace(QUOTE, ESCAPED_QUOTE);

        if (s.IndexOfAny(CHARACTERS_THAT_MUST_BE_QUOTED) > -1)
            s = QUOTE + s + QUOTE;

        return s;
    }

在这种情况下，我不确定在哪里可以使用我的转义函数。

示例：

学位，研究生，08年12月8日，第1级，“高级，数学”

字符串Advanced，Maths被分成两个不同的数组项，我不想

Answer 1

不确定这是如何实现的 - 但您可以使用Linq.Aggregate来解决这个问题：

using System;
using System.Linq;
using System.Collections.Generic;

public class Program
{
    public static IEnumerable<string> SplitIt(
        char[] splitters, 
        string text, 
        StringSplitOptions opt = StringSplitOptions.None)
    {
        bool inside = false;
        var result = text.Aggregate(new List<string>(), (acc, c) =>
        {
            // this will check each char of your given text
            // and accumulate it in the (empty starting) string list
            // your splitting chars will lead to a new item put into 
            // the list if they are not inside. inside starst as false
            // and is flipped anytime it hits a "
            // at end we either return all that was parsed or only those
            // that are neither null nor "" depending on given opt's
            if (!acc.Any()) // nothing in yet
            {
                if (c != '"' && (!splitters.Contains(c) || inside))
                    acc.Add("" + c);

                else if (c == '"')
                    inside = !inside;

                else if (!inside && splitters.Contains(c)) // ",bla"
                    acc.Add(null);

                return acc;
            }

            if (c != '"' && (!splitters.Contains(c) || inside))
                acc[acc.Count - 1] = (acc[acc.Count - 1] ?? "") + c;

            else if (c == '"')
                inside = !inside;

            else if (!inside && splitters.Contains(c)) // ",bla"
                acc.Add(null);

            return acc;
        }

        );
        if (opt == StringSplitOptions.RemoveEmptyEntries)
            return result.Where(r => !string.IsNullOrEmpty(r));

        return result;
    }

    public static void Main()
    {
        var s = ",,Degree,Graduate,08-Dec-17,Level 1,\"Advanced, Maths\",,";
        var spl = SplitIt(new[]{','}, s);
        var spl2 = SplitIt(new[]{','}, s, StringSplitOptions.RemoveEmptyEntries);
        Console.WriteLine(string.Join("|", spl));
        Console.WriteLine(string.Join("|", spl2));
    }
}

输出：

|Degree|Graduate|08-Dec-17|Level 1|Advanced, Maths||
Degree|Graduate|08-Dec-17|Level 1|Advanced, Maths

Answer 2

您可以使用正则表达式，linq或仅循环遍历每个字符，并使用布尔值来确定当前的行为应该是什么。这个问题实际上让我思考，因为我之前只是循环并对每个角色采取行动。这是Linq打破整个csv文档的方法，假设可以在';'找到行尾：

private static void Main(string[] args)
    {
        string example = "\"Hello World, My name is Gumpy!\",20,male;My sister's name is Amy,29,female";

        var result1 = example.Split(';')
                             .Select(s => s.Split('"')) // This will leave anything in abbreviation marks at odd numbers
                             .Select(sl => sl.Select((ss, index) => index % 2 == 0 ? ss.Split(',') : new string[] { ss })) // if it's an even number split by a comma
                             .Select(sl => sl.SelectMany(sc => sc));

        Console.WriteLine("Press any key to continue.");
        Console.ReadKey();
    }

Answer 3

该函数在字符串中获取逗号分隔的字段，不包括嵌入在引用字段中的逗号

假设

应返回空字段,,
引号字段中没有引号（根据示例）

方法

我使用for loop i作为当前字段的占位符
它会扫描下一个逗号或引号，如果找到引号，则会扫描下一个逗号以创建字段
它需要高效，否则我们会使用正则表达式或Linq
OP并不想使用CSV库

注意：没有错误检查，扫描每个字符会更快，这很容易理解

代码

public List<string> GetFields(string line)
{
    var list = new List<string>();

    for (var i = 0; i < line.Length; i++)
    {
        var firstQuote = line.IndexOf('"', i);
        var firstComma = line.IndexOf(',', i);

        if (firstComma >= 0)
        {
            // first comma is before the first quote, then its just a standard field
            if (firstComma < firstQuote || firstQuote == -1)
            {
                list.Add(line.Substring(i, firstComma - i));
                i = firstComma;
                continue;
            }

            // We have found quote so look for the next comma afterwards
            var nextQuote = line.IndexOf('"', firstQuote + 1);
            var nextComma = line.IndexOf(',', nextQuote + 1);

            // if we found a comma, then we have found the end of this field
            if (nextComma >= 0)
            {
                list.Add(line.Substring(i, nextComma - i));
                i = nextComma;
                continue;
            }
        }

        list.Add(line.Substring(i)); // if were are here there are no more fields
        break;

    }
    return list;
}

测试1

学位，研究生，08年12月18日，1级，＆＃34;高级，数学＆＃34;，另一个

Degree
Graduate
08-Dec-17
Level 1
"Advanced, Maths"
another

测试2

，学位，研究生，08年12月8日，\＆＃34; asdasd \＆＃34;，1级，\＆＃34;高级，数学\＆＃34;，另一个

<Empty Line>
Degree
Graduate
08-Dec-17
"asdasd"
Level 1
"Advanced, Maths"
another

不要通过转义字符串拆分 - C＃

3 个答案: