為什么 for 循環在給定的 Java 代碼中花費太多時間

Question

如何減少 for 循環所花費的時間，如果可能，使用 replaceAll(,) 方法刪除 for 循環：

String extractText(String s) throws IOException
{
    String html = fj.toHtmlString(s); //extracted html source code from wikipedia
    String filtered_text="";
    System.out.println("extracted \n\n");
    String []html_text = html.split("\n");
    long start = System.currentTimeMillis();

    for(String h:html_text)
    {   //System.out.println("ky4"+h);
        if(Pattern.compile("</strong>", Pattern.CASE_INSENSITIVE + Pattern.LITERAL).matcher(h).find())
        {

        }
        else if(Pattern.compile("<strong", Pattern.CASE_INSENSITIVE + Pattern.LITERAL).matcher(h).find())
        {

        }
        else
        {
            filtered_text += h;
            filtered_text += "\n";
        }
    }
    long end = System.currentTimeMillis();
    System.out.println("loop end in "+(end-start)/1000+" seconds"+" or "+(end-start)+" miliseconds");//System.out.println(++i2+" th loop end in "+(end-start)/1000+" seconds");
    return filtered_text;
}

Answer 1

讓我們來看看循環：

for(String h:html_text)
{   //System.out.println("ky4"+h);
    if(Pattern.compile("</strong>", Pattern.CASE_INSENSITIVE + Pattern.LITERAL).matcher(h).find())
    {

    }
    else if(Pattern.compile("<strong", Pattern.CASE_INSENSITIVE + Pattern.LITERAL).matcher(h).find())
    {

    }
    else
    {
        filtered_text += h;
        filtered_text += "\n";
    }
}

每次循環至少要編譯一個正則表達式，這非常耗時。 改用變量：

Pattern endingTag = Pattern.compile("</strong>", Pattern.CASE_INSENSITIVE + Pattern.LITERAL);
Pattern startTag = Pattern.compile("<strong", Pattern.CASE_INSENSITIVE + Pattern.LITERAL)
for(String h:html_text)
{   //System.out.println("ky4"+h);
    if(endingTag.matcher(h).find())
    {}
    else if(startTag.matcher(h).find())
    {}
    else
    {
        filtered_text += h;
        filtered_text += "\n";
    }
}

這將為您節省大量時間。 另請注意，在測量性能時，請始終在發布模式下運行。

Answer 2

您可以嘗試在循環之前而不是在循環中編譯您的模式：

String extractText(String s) throws IOException
{
    String html = fj.toHtmlString(s); //extracted html source code from wikipedia
    String filtered_text="";
    System.out.println("extracted \n\n");
    String []html_text = html.split("\n");
    long start = System.currentTimeMillis();
    Pattern end = Pattern.compile("</strong>", Pattern.CASE_INSENSITIVE + Pattern.LITERAL)
    Pattern start = Pattern.compile("<strong", Pattern.CASE_INSENSITIVE + Pattern.LITERAL)
    for(String h:html_text)

Answer 3

我的問題通過以下代碼解決：

String extractText(String s) throws IOException
{
    String html = fj.toHtmlString(s); 
    String filtered_text="";
    System.out.println("extracted \n\n");
    html=html.replaceAll("(?i)</strong>", "");
    html=html.replaceAll("(?i)<strong[^>]*>", "");
    filtered_text = html;        
    long end = System.currentTimeMillis();
    System.out.println("loop end in "+(end-start)/1000+" seconds"+" or "+(end-start)+" miliseconds");//System.out.println(++i2+" th loop end in "+(end-start)/1000+" seconds");
    return filtered_text;
}