C#快速随机按行读取大型文本文件

原文:C#快速随机按行读取大型文本文件

下面是我实现的一个数据文件随机读取类,可以随机读取大型文本文件的某一行。在我机器上对一个130MB的文本文件,读取第200000的速度从传统做法的400ms提高到了3ms。
一般对文本文件进行读取时,一般采用ReadLine()进行逐行读取。在这种情况下,C#内的FileStream和BufferedStream类处理绰绰有余了。它不会将整个文件全部读入,而是有缓冲的读。但是,要想随机读取某一行,在行数据长度不统一的情况下,如果每次这样遍历到指定行,其效率显然是很低下的。
当然,代价也是有的,引入了第一次打开文件的打开时间,且占用了少部分内存(占用多少是可以设置的,当然占得越小速度也越慢,但最大值也比全部读入要小很多)。

(对网络代码进行部分改写)

using System;
using System.Collections.Generic;
using System.Text;
using System.Collections;
using System.Threading;
using System.IO;

namespace DataBuffer
{
    public static class FileConfig
    {
        public static int STREAM_BUFFER_SIZE = 1024000;
        public static int MAP_DISTANCE = 10;
    }

    public class DataFile
    {
        ///
        /// 数据文件名
        ///
        public string fileName = "";
        ///
        /// 初始化读取完标志
        ///
        public bool done = false;

        ///
        /// 当前流位置
        ///
        public long Position = 0;

        ///
        /// 文件头部信息
        ///
        private Hashtable head = new Hashtable();
        public Hashtable Head { get { return head; } set { head = value; } }

        ///
        /// 文件地图
        ///       
        private ArrayList map = new ArrayList();
        public ArrayList Map { get { return map; } set { map = value; } }

        ///
        /// 文件数据行行数
        ///       
        private long lines = 0;
        public long Lines { get { return lines; } set { lines = value; } }
    }

    public class DataBuffer
    {
        private FileStream fs = null;
        private BufferedStream bs = null;
        private StreamReader sr = null;
        private StreamWriter sw = null;
        ///
        /// 文件信息数据结构
        ///
        public DataFile dataFile = new DataFile();

        public DataBuffer(string name)
        {
            dataFile.fileName = name;
        }

        ///
        /// 打开文件
        ///
        public bool Open()
        {
            try
            {
                //初始化各流
                fs = new FileStream(dataFile.fileName, FileMode.Open, FileAccess.ReadWrite);
                bs = new BufferedStream(fs, FileConfig.STREAM_BUFFER_SIZE);
                sr = new StreamReader(fs);
                sw = new StreamWriter(fs);
                Thread initFile = new Thread(new ThreadStart(InitDataFile));
                initFile.Start();
                return true;
            }
            catch (Exception ee)
            {
                ErrorHandler.ErrorHandler eh = new ErrorHandler.ErrorHandler(ee, "文件打开");
                return false;
            }
        }

        private void InitDataFile()
        {
            //另开一个读取流
            BufferedStream bs = new BufferedStream(fs);
            StreamReader sr = new StreamReader(bs);

            //读入数据文件头信息。共14行
            string thisLine = NextLine(ref sr);
            dataFile.Head.Add("Subject", thisLine.Substring(11));

            thisLine = NextLine(ref sr);
            dataFile.Head.Add("Date", thisLine.Substring(8));

            thisLine = NextLine(ref sr);
            dataFile.Head.Add("Time", thisLine.Substring(8));

            thisLine = NextLine(ref sr);
            dataFile.Head.Add("Channels", thisLine.Substring(12));

            thisLine = NextLine(ref sr);
            dataFile.Head.Add("Rate", thisLine.Substring(8));

            thisLine = NextLine(ref sr);
            dataFile.Head.Add("Type", thisLine.Substring(8));

            thisLine = NextLine(ref sr);
            dataFile.Head.Add("Rows", thisLine.Substring(8));

            thisLine = NextLine(ref sr);
            thisLine = NextLine(ref sr);
            dataFile.Head.Add("Electrode Labels", thisLine);
            thisLine = NextLine(ref sr);
            thisLine = NextLine(ref sr);
            thisLine = NextLine(ref sr);
            thisLine = NextLine(ref sr);
            thisLine = NextLine(ref sr);
            //降低自己的优先级
            Thread.CurrentThread.Priority = ThreadPriority.BelowNormal;

            //数行数,建立地图
            long lines = 1;
            //在地图中加入首条数据的位置信息
            dataFile.Map.Add(dataFile.Position);
            //顺序建立文件地图
            while (!sr.EndOfStream)
            {
                thisLine = NextLine(ref sr);
                if ((++lines) % FileConfig.MAP_DISTANCE == 0)
                {
                    dataFile.Map.Add(dataFile.Position);
                }
            }
            dataFile.Lines = lines;
            dataFile.done = true;
        }

        ///
        /// 文件关闭
        ///
        public bool Close()
        {
            try
            {
                //顺序关闭各流
                sw.Close();
                sr.Close();
                bs.Close();
                fs.Close();
                return true;
            }
            catch (Exception ee)
            {
                ErrorHandler.ErrorHandler eh = new ErrorHandler.ErrorHandler(ee, "文件关闭");
                return false;
            }
        }

        ///
        /// 顺序读取下一行。效率低不建议大规模使用,只在打开文件的时候使用一次
        ///
        ///
        public string NextLine(ref StreamReader sr)
        {
            string next = sr.ReadLine();
            //+2是指Windows换行回车。Linux下要改为+1
            dataFile.Position += next.Length + 2;
            return next;
        }

        //指定的目标行内容
        public string ReadLine(long line)
        {
            try
            {
                //如果载入完毕
                if (dataFile.done)
                {
                    //确定数据块索引号
                    int index = (int)line / FileConfig.MAP_DISTANCE;
                    //移动到指定位置
                    bs.Seek(long.Parse(dataFile.Map[index].ToString()), SeekOrigin.Begin);
                    //创建流读取器
                    sr = new StreamReader(bs);
                    //移动到指定行
                    for (int i = 1; i <= (line - index * FileConfig.MAP_DISTANCE); i++)
                    {
                        sr.ReadLine();
                    }
                    //返回指定行的值
                    return sr.ReadLine();
                }
                else
                {
                    return "";
                }
            }
            catch (Exception ee)
            {
                ErrorHandler.ErrorHandler eh = new ErrorHandler.ErrorHandler(ee, "文件读取");
                return "";
            }
        }
    }

}

C#快速随机按行读取大型文本文件

上一篇:平方根倒数快速计算


下一篇:AtCoder Beginner Contest 214