字符串线性搜索c#

Question

我目前正在做一个小型 C# 练习，处理在文本文件中搜索相关术语/单词，程序将写出文本文件中包含搜索单词的所有句子。例如，我输入单词：“example”，程序将遍历文本文件中的所有句子，并提取其中包含“example”一词的句子。

The text file is structured as so: <sentenceDesignator> <text>
sentence 1: bla bla bla bla example of a sentence  //each line contains a sentence
sentence 2: this is not a good example of grammar
sentence 3: bla is not a real word, use better terms

我希望能够做的是使用线性搜索来遍历文本文件中的所有行并写出包含搜索到的字符串术语的所有句子。

到目前为止我的代码：

        String filename = @"sentences.txt";

        if (!File.Exists(filename))
        {
            // Since we just created the file, this shouldn't happen.
            Console.WriteLine("{0} not found", filename);
            return;
        }
        else
        {
            Console.WriteLine("Successfully found {0}.", filename);
        }
        //making a listof type "Sentence" to hold all the sentences
        List<Sentence> sentences = new List<Sentence>();

        //the next lines of code...
        StreamReader reader = File.OpenText(filename);

        //first, write out all of the sentences in the text file

        //read a line(sentence) from a line in the text file
        string line = reader.ReadLine();

        while (line != null)
        {
            Sentence s = new Sentence();

            //we need something to split data...
            string[] lineArray = line.Split(':');

            s.sentenceDesignator = lineArray[0];
            s.Text = lineArray[1];

            Console.Write("\n{0}", line);

            line = reader.ReadLine();
        }

        //so far, we can write out all of the sentences in the text file. 
        Console.Write("\n\nOK!, search a term to diplay all their occurences: ");
        string searchTerm = Console.ReadLine();

       if(!line.Contains(searchterm))
       {
          Console.Write("\nThat term does not exist in any sentence.");
       }
       else
        {
            foreach (Sentence ss in sentences)
            {
                if (ss.sentenceDesignator.Contains(queryName))
                {
                    //I need help here
                }
            }
        }

Answer 1

如果您构建文件的索引然后搜索索引会快很多，因为对于线性搜索，每个搜索操作都是 O(n) 而对于索引搜索，它是 O(n) 用于构建索引，但是O(log n) 或 near-O(1) 用于查找（取决于您如何构建索引）。代价是增加了索引的内存消耗，但我会这样做：

private Dictionary<String,List<Int32>> _index = new Dictionary<String,List<Int32>>();

/// <summary>Populates an index of words in a text file. Takes O(n) where n is the size of the input text file.</summary>
public void BuildIndex(String fileName) {

    using(Stream inputTextFile = OpenFile(...)) {

        int currentPosition = 0;
        foreach(String word in GetWords(inputTextFile)) {

            word = word.ToUpperInvariant();
            if( !_index.ContainsKey( word ) ) _index.Add( word, new List<Int32>() );
            _index[word].Add( currentPosition );

            currentPosition = inputTextFile.Position;
        }
    }
}

/// <summary>Searches the text file (via its index) if the specified string (in its entirety) exists in the document. If so, it returns the position in the document where the string starts. Otherwise it returns -1. Lookup time is O(1) on the size of the input text file, and O(n) for the length of the query string.</summary>
public Int32 SearchIndex(String query) {

    String[] terms = query.Split(' ');

    Int32 startingPosition = -1;
    Int32 currentPosition = -1;
    Boolean first = true;
    foreach(String term in terms) {
        term = term.ToUpperInvariant();

        if( first ) {
            if( !_index.Contains( term ) ) return -1;
            startingPosition = _index[term][0];
        } else {

            if( !ContainsTerm( term, ++currentPosition ) ) return -1;
        }

        first = false;
    }

    return startingPosition;
}

/// <summary>Indicates if the specified term exists at the specified position.</summary>
private Boolean ContainsTerm(String term, Int32 expectedPosition) {

    if( !_index.ContainsKey(term) ) return false;
    List<Int32> positions = _index[term];
    foreach(Int32 pos in positions) {

        if( pos == expectedPosition ) return true;
    }
    return false;
}

OpenFile 和 GetWords 的实现应该是微不足道的。请注意，GetWords 使用 yield return 在文件中构建由空格分隔的单词的 IEnumerable<String>，以及处理您的自定义文件格式。

Answer 2

我对最后一个 if/else 有点困惑。似乎您只是将文件的最后一行与搜索词进行比较。另外，“queryName”从何而来？你想打印出整个句子（“bla bla bla bla example of a sentence”）还是只打印“sentence 1”？另外，您检查 sentenceDesignator 是否包含 queryName，我想您想检查实际的 Text 是否包含搜索词。

也许这会对你有所帮助：

var lines = File.ReadAllLines(fileName);    
var sentences = new List<Sentence>(lines.Count());

foreach (var line in lines)
{
    var lineArray = line.Split(':');
    sentences.Add(new Sentence { sentenceDesignator = lineArray[0], Text = lineArray[1]});
}

foreach (var sentence in sentences)
{
    if (sentence.Text.Contains(searchTerm))
    {
        Console.WriteLine(sentence.sentenceDesignator);
        //Console.WriteLine(sentence.Text);
    }
}