StreamReader 跳过 Null 或 WhiteSpace 行的最佳方式

Question

在搜索并尝试了不同的方法后，我发现我要么对我编写代码的方式不满意，要么它不适合我。我是编程新手，所以我的理解有限。请记住答案。

我想逐行读取 .csv 文件并跳过空白行。将行的内容放入对象列表中。除了跳线部分，我一切正常。也欢迎任何关于改进我的代码任何部分的反馈。我喜欢建设性的批评。

public void CardaxCsvFileReader()
    {
        string cardaxCsvPath = (@"C:\Cardax2WkbTest\Cardax\CardaxTable.csv");

        try
        {
            using (System.IO.StreamReader cardaxSR =
                new System.IO.StreamReader(System.IO.File.OpenRead(cardaxCsvPath)))
            {
                string line = "";
                string[] value = line.Split(',');

                while (!cardaxSR.EndOfStream)
                {                              // this commented out part is what I would like to work but doesn't seem to work.
                    line = cardaxSR.ReadLine();//.Skip(1).Where(item => !String.IsNullOrWhiteSpace(item));
                    value = line.Split(',');

                    if (line != ",,,,,") // using this as temp to skip the line because the above commented out part doesn't work.
                    {
                    CardaxDataObject cardaxCsvTest2 = new CardaxDataObject();

                    cardaxCsvTest2.EventID = Convert.ToInt32(value[0]);
                    cardaxCsvTest2.FTItemID = Convert.ToInt32(value[1]);
                    cardaxCsvTest2.PayrollNumber = Convert.ToInt32(value[2]);
                    cardaxCsvTest2.EventDateTime = Convert.ToDateTime(value[3]);
                    cardaxCsvTest2.CardholderFirstName = value[4];
                    cardaxCsvTest2.CardholderLastName = value[5];

                    Globals.CardaxQueryResult.Add(cardaxCsvTest2);
                    }
                }
            }
        }
        catch (Exception)
        {
            myLog.Error("Unable to open/read Cardax simulated punch csv file! " +
                "File already open or does not exist: \"{0}\"", cardaxCsvPath);
        }

Answer 1

已编辑

如果您的行不是真正的空白并且包含逗号，您可以使用RemoveEmptyEntries 选项进行拆分，然后检查列数。

                while (!cardaxSR.EndOfStream)
                {                              // this commented out part is what I would like to work but doesn't seem to work.
                    line = cardaxSR.ReadLine();//.Skip(1).Where(item => !String.IsNullOrWhiteSpace(item));
                    value = line.Split(new char[] {','}, StringSplitOptions.RemoveEmptyEntries);  // <-- Remove empty columns while splitting. It has a side-effect: Any record with just a single blank column will also get discarded by the if that follows.
                    if (value.length < 6)
                      continue;

                    CardaxDataObject cardaxCsvTest2 = new CardaxDataObject();

                    cardaxCsvTest2.EventID = Convert.ToInt32(value[0]);
                    cardaxCsvTest2.FTItemID = Convert.ToInt32(value[1]);
                    cardaxCsvTest2.PayrollNumber = Convert.ToInt32(value[2]);
                    cardaxCsvTest2.EventDateTime = Convert.ToDateTime(value[3]);
                    cardaxCsvTest2.CardholderFirstName = value[4];
                    cardaxCsvTest2.CardholderLastName = value[5];

                    Globals.CardaxQueryResult.Add(cardaxCsvTest2);
                }

我有另一个改进反馈：当您捕获异常时，除了您的自定义错误行之外，记录该异常是一个好习惯。自定义错误行可能对网站用户有好处，但作为运行某些服务的开发人员，您将欣赏实际的异常堆栈跟踪。它将帮助您更轻松地调试错误。

    catch (Exception ex)
    {
        myLog.Error("Unable to open/read Cardax simulated punch csv file! " +
            "File already open or does not exist: \"{0}\".\r\n Exception: {1}", cardaxCsvPath, ex.ToString());
    }

Answer 2

只需检查value.Length == 6，这样它就会跳过不包含足够数据的行

Answer 3

使用专用 CSV 解析器，例如此处提供的 EasyCSV 类^*：

https://github.com/jcoehoorn/EasyCSV

public void CardaxCsvFileReader()
{
    try 
    {
        string cardaxCsvPath = (@"C:\Cardax2WkbTest\Cardax\CardaxTable.csv");
        Globals.CardaxQueryResult =
           EasyCSV.FromFile(cardaxCsvPath)
              .Where(r => r.Any(c => !string.IsNullOrEmpty(c)))
              .Select(r => CardaxDataObject() {
                    cardaxCsvTest2.EventID = int.Parse(r[0]),
                    cardaxCsvTest2.FTItemID = int.Parse(r[1]),
                    cardaxCsvTest2.PayrollNumber = int.Parse(r[2]),
                    cardaxCsvTest2.EventDateTime = DateTinme.Parse(r[3]),
                    cardaxCsvTest2.CardholderFirstName = r[4],
                    cardaxCsvTest2.CardholderLastName = r[5]
              }).ToList();      
    }
    catch (Exception)
    {
        myLog.Error("Unable to open/read Cardax simulated punch csv file! " +
            "File already open or does not exist: \"{0}\"", cardaxCsvPath);
    }
}

我还建议您重新考虑如何构建它。下面的代码是更好的做法：

public IEnumerable<CardaxDataObject> ReadCardaxCsvFile(string filename)
{
   //no try block at this level. Catch that in the method that calls this method

    return EasyCSV.FromFile(cardaxCsvPath)
          .Where(r => r.Any(c => !string.IsNullOrEmpty(c)))
          // You may want to put a try/catch inside the `Select()` projection, though.
          // It would allow you continue if you fail to parse an individual record
          .Select(r => CardaxDataObject() {
                cardaxCsvTest2.EventID = int.Parse(r[0]),
                cardaxCsvTest2.FTItemID = int.Parse(r[1]),
                cardaxCsvTest2.PayrollNumber = int.Parse(r[2]),
                cardaxCsvTest2.EventDateTime = DateTinme.Parse(r[3]),
                cardaxCsvTest2.CardholderFirstName = r[4],
                cardaxCsvTest2.CardholderLastName = r[5]
         });      
}

突然间，该方法归结为一个语句（尽管语句很长）。像这样的代码更好，因为它更强大，原因有三个：它不限于只使用一个输入文件，它不限于只将其输出发送到一个位置，它不限于只有一种处理错误的方法。你可以这样称呼它：

try 
{
    string cardaxCsvPath = (@"C:\Cardax2WkbTest\Cardax\CardaxTable.csv");
    Globals.CardaxQueryResult = ReadCardaxCsvFile(cardaxCsvPath).ToList();
}
catch (Exception)
{
    myLog.Error("Unable to open/read Cardax simulated punch csv file! " +
            "File already open or does not exist: \"{0}\"", cardaxCsvPath);
}

或者像这样：

try 
{
    string cardaxCsvPath = (@"C:\Cardax2WkbTest\Cardax\CardaxTable.csv");
    foreach (var result in ReadCardaxCsvFile(cardaxCsvPath))
    {
        Globals.CardaxQueryResult.Add(result);
    }
}
catch (Exception)
{
    myLog.Error("Unable to open/read Cardax simulated punch csv file! " +
            "File already open or does not exist: \"{0}\"", cardaxCsvPath);
}

我也建议不要使用这样的Globalsclass。找到一个更有意义的对象，您可以将其与此数据相关联。

---

^{* 免责声明：我是该解析器的作者}