假设我有一个输入字符串:
String test = "\"item one\",\"item,2\",12345";
String[] arr = test.Split(',');
结果是:
[0]: "item one"
[1]: "item"
[2]: "2"
[3]: "12345"
但我想:
[0]: "item one"
[1]: "item, 2"
[2]: "12345"
基本上,我的输入字符串将有引号,以允许引号之间使用逗号,这不是分隔符。
将此行解析为数组的最佳方法是什么?
答案 0 :(得分:3)
下面是简单而完整的CsvHelper,它实现了从/向csv数据线编码解码线。在我的上一个项目中,我想使用FastCSVReader,但内部看起来相当复杂,因为我没有尽可能快速和最强大的解决方案,我写了我的小助手。
/// <summary>
/// CsvHelper
/// </summary>
public static class CsvHelper
{
#region Public methods
/// <summary>
/// Codes fields as a line for csv file
/// </summary>
/// <param name="fields"></param>
/// <returns></returns>
public static string CodeLine(string[] fields)
{
if (fields == null || fields.Length == 0)
return null;
var sb = new StringBuilder(1024);
for (int i = 0; i < fields.Length; i++)
{
if (fields[i] != null)
{
bool mustBeQuoted = fields[i].Contains("\"") || fields[i].Contains(",");
if (mustBeQuoted)
{
sb.Append("\"");
sb.Append(fields[i].Replace("\"", "\"\""));
sb.Append("\"");
}
else
{
sb.Append(fields[i]);
}
}
if (i != fields.Length - 1)
sb.Append(",");
}
return sb.ToString();
}
/// <summary>
/// Decodes line from csv file into fields
/// </summary>
/// <param name="line"></param>
/// <param name="fields"></param>
/// <returns>Returns true if decoding was successful</returns>
public static bool DecodeLine(string line, out string[] fields)
{
fields = null;
if (string.IsNullOrEmpty(line))
return false;
int index = 0;
var res = new List<string>();
while (index != line.Length)
{
string field;
if (ReadField(line, ref index, out field))
{
res.Add(field);
}
else
{
return false;
}
}
if (line[line.Length - 1] == ',')
{
res.Add(string.Empty);
}
fields = res.ToArray();
return true;
}
#endregion
#region Other methods
private static bool ReadField(string line, ref int index, out string field)
{
field = null;
if (index >= line.Length)
return false;
var sb = new StringBuilder(512);
int state = 0;
while (true)
{
char c = line[index];
char? c1 = (index + 1 < line.Length - 1) ? (char?)line[index + 1] : null;
index++;
switch (state)
{
case 0: // START
if (c == '"') //text in field starts by quotation mark, text of filed in quotation marks
{
state = 4;
}
else if (c == ',') //empty text of filed
{
field = sb.ToString();
return true;
}
else
{
state = 1;
sb.Append(c);
}
break;
case 1: //Not quoted text in field
if (c == '"') // error, cannot contains " in the middle of the field
{
return false;
}
else if (c == ',')
{
field = sb.ToString();
return true;
}
else
{
sb.Append(c);
}
break;
case 3: //Escaping quotation mark
if (c == '"') //previous quotation mark was escape char for this quotation mark
{
state = 4;
sb.Append(c);
}
else //error, cannot contais any other char
{
return false;
}
break;
case 4: //Text in between quotation marks
if (c == '"') //closing quoted text or escape char for following qoatation mark - based on which char is following
{
if (c1 != null && c1.Value == '"') //current quotation mark is escape char for following quotation mark
{
state = 3;
}
else
{
state = 5;
}
}
else
{
sb.Append(c);
}
break;
case 5: //Just after closing quotation mark of quoted text
if (c == ',') //closing quoted text
{
field = sb.ToString();
return true;
}
else //error, cannot contais any other char
{
return false;
}
break;
}
if (index == line.Length)
{
if (state == 1 || state == 5)
{
field = sb.ToString();
return true;
}
return false;
}
}
}
#endregion
}
答案 1 :(得分:0)
我可能会像(["])[^"]*\1|(\d+)
一样使用正则表达式,但当然可能需要考虑使用这种方法的其他格式。