在一段时间内不断请求和收集数据

Question

我需要通过发送命令然后读取结果，在 X 分钟内尽可能快地连续检索数据。我不确定以下代码是否有效或用于不断检索：

DateTime utcStartTime = DateTime.UtcNow;
        while (DateTime.UtcNow <= (utcStartTime.AddSeconds(recordTime)))                       //Run the read-write function for approximately the time specified
        {
            try
            {

                Write("Write.Request");                                 //Requests the data
                Pmeas = Read();                                         //Reads the returned string. String = "ReadError" if no result is found.
                                                                        //Note that the following error checker doesn't effect the speed of data collection at a millisecond level (already checked), and is therefore not the cause of any delay.
                if (String.IsNullOrEmpty(Pmeas))
                {
                    Pmeas = "ReadError";                                //Necessary due to an apparent overwrite caused by the read function - Tells the later output (outside of while loop) that Pmeas experienced a read error
                    DateTime utcTime = DateTime.UtcNow;                 //Get the current time in UTC seconds (may need correcting).
                    pArray[i] = (Pmeas + "\t" + utcTime.ToString("%d") + Environment.NewLine);    //Appends the Pmeas of each instance to a string array with a timestamp
                }
                else
                {
                    DateTime utcTime = DateTime.UtcNow;
                    pArray[i] = (Pmeas + "\t" + utcTime.ToString("%d") + Environment.NewLine);    //Appends the Pmeas of each instance to a string array with a timestamp
                }
                Pmeas = "ReadError";                                    //Reset Pmeas to prove in file that Pmeas experienced a read error
            }
            catch (Exception f)                                         //Catch an exception if try fails
            {
                Console.WriteLine("{0} Exception caught.", f);
            }
            i++;                                            //let i grow so that the array can also grow, plus have a variable already available for being having the string being written into a file (not part of question or code-in-question).
        }

请注意，所有变量，如 i、Pmeas 和 pArray 都是在循环之前预定义的（以防止错误），并且 UTC 应该以秒为单位（不确定当前的语法；具有十进制精度）。正如我所说，我正在寻找一种方法，该方法可以使用上面给出的读取和写入函数不断地从另一个来源收集数据，并在一段时间内不断地这样做而不会延迟。我的简单问题是，这是收集数据的正确方法还是有更好和/或更有效的方法？

欢迎对代码的所有输入，即使它没有完整回答问题。

Answer 1

字符串连接可能很慢，请尝试创建一个包含 DateTime 和字符串的类，并让 pArray 代替它。稍后在时间不重要时转换为字符串。也可以使用秒表来记录时长，DateTime 的最小分辨率约为 15ms。

//Declared elsewhere
public class DataPoint
{
    public TimeSpan Time {get; set;}
    public String Message {get; set;}
}


    List<DataPoint> dataPoints = new List<DataPoint>(pArray.Length); //make the default size the same as pArray so we don't waist time growing the list.

    Stopwatch duration = new Stopwatch();
    DateTime utcStartTime = DateTime.UtcNow;
    duration.Start();
    DateTime endTime = utcStartTime.AddSeconds(recordTime); //Move this out of the loop so it only does the calculation once.
    while (DateTime.UtcNow <= endTime)                       //Run the read-write function for approximately the time specified
    {
        try
        {

            Write("Write.Request");                                 //Requests the data
            Pmeas = Read();                                         //Reads the returned string. String = "ReadError" if no result is found.

            var dataPoint = new DataPoint
                {
                    Time = duration.Elapsed,
                    Message = Pmeas
                };
            dataPoints.Add(dataPoint);                                                     
            Pmeas = "ReadError";                                    //Reset Pmeas to prove in file that Pmeas experienced a read error
        }
        catch (Exception f)                                         //Catch an exception if try fails
        {
            Console.WriteLine("{0} Exception caught.", f);
        }
    }

    //Now that we are out of the time critical section do the slow work of formatting the data.
    foreach(var dataPoint in dataPoints)
    {
        var message = dataPoint.Message;
        if (String.IsNullOrEmpty(message))
        {
            message = "ReadError";
        }
        pArray[i] = message + "\t" + (utcStartTime + dataPoint.Time).ToString("%d") + Environment.NewLine;    //Appends the Pmeas of each instance to a string array with a timestamp

        i++;                                            //let i grow so that the array can also grow, plus have a variable already available for being having the string being written into a file (not part of question or code-in-question).
    }

但这可能是一个微小的变化，请获取一个分析器（visual studio comes with one，但我更喜欢使用DotTrace），看看实际花费最多的时间并集中精力在那里。