如何在 C++ 中正确获取 input/output 流的 file/stream 位置指针?

How to correctly get the file/stream position pointers in C++ for an input/output stream?

如果我有一个std::stringstream变量并且我在里面写了一些std::string objects,我希望能够跟踪流位置指针并将它们输出到屏幕上,我该如何使用tellp() 然后当我从流中提取时,即使用流作为输入,我如何使用 tellg()。 如何再次以输入模式打开流?另外,当我再次打开它输入时,我应该用 seekg()seekp() 将位置指针设置回 0 吗?

这是我的代码

std::string word;
std::stringstream ss;

while (word != "n") {
  std::cout << "Enter a word, ( n to stop) : ";
  std::cin >> word;
  std::cin.ignore();
  if (word == "n")
    ss << std::endl;
  else {
    ss << word;
    ss << " ";
  }
}

std::cout << "The content in stringstream is : " << ss.str() << std::endl;

std::cout << "The StreamPositionPointer is"
          << ss.tellp();  // is this correct way of outputting streampointer

// how should I open the stream to get input from it to store it in a variable?

/* how should I switch the position pointer to a position when I use an
   input/output stream like sstream with seekg(), seekp(), tellg(),tellp() */
std::string word;
std::stringstream ss;

while (word != "n") {
  std::cout << "Enter a word, ( n to stop) : ";
  std::cin >> word;
  std::cin.ignore();
  if (word == "n")
    ss << std::endl;
  else {
    ss << word;
    ss << " ";
  }
}


std::cout << "The content in stringstream is : " << ss.str() << std::endl;

std::cout << "The StreamPositionPosition is: "
          << ss.tellp() << std::endl;  // Yes this is correct, it will give you the position of the current position where you are outputting.

// how should i open the stream to get input from it to store it in a variable?
// If you want to convert it to string simply
std::string str = ss.str();
// If you know the order of the data, you can simply extract it
std::string wordOutput;
ss >> wordOutput;

/* how should I switch the position pointer to a position when I use an
   input/output stream like sstream with seekg(), seekp(), tellg(),tellp() */
/*
seekg sets the position of input
tellg gets the position of input

seekp sets the position of output
tellp gets the position of output
*/

long length = s.tellp(); // gives you the length of the string stream
// If you want the first position do,
 ss.seekp(0);
ss >> wordOutput;
// If you want the third last, do
ss.seekp(length - 3);
ss >> wordOutput;

// Examples for inputs
stringstream example;
example.write ("This is an apple",16);
long pos = example.tellp();
example.seekp (pos-7);
example.write (" sam",4);
// Will result in This is a sample

你有基本的想法。 std::basic_stringstream allows use of both std::basic_istream::tellg and std::basic_istream::seekg。有几点注意事项:

  1. 如果使用 .str() 成员函数,底层字符串的副本被 returned 在一个临时对象中,stringstream 中的 seekpos 没有变化,因为你正在操作一个复制;
  2. 如果使用rdbuf()成员函数,它将return一个指向底层字符串设备的指针。 output with the device 会输出with的内容,seekpos会前进到最后,但是eofbit没有设置;
  3. 对于使用 >> 读取字符串流集合中最后一项的正常提取 eofbitclear() 必须在进一步查找之前在字符串流上调用。如果在 stringstream 时调用 tellg(),streamstate 不是 good-1 是 returned.

通过一个简短的例子,练习 seekg()tellg() 以及 clear(),在需要的地方,应该可以提供您正在寻找的解释。这是上面链接中提供的各种示例的混合体:

#include <iostream>
#include <iomanip>
#include <string>
#include <sstream>
 
int main (void)
{
    std::string str = "Hello, world";         /* initial string */
    std::istringstream in(str);               /* initialized stringstream */
    std::string word1, word2;                 /* two string variables used below */
    
    std::cout << "string: " << in.str() << "\n\ntellg " 
                << std::setw(2) << in.tellg() << " - reading into word1\n";
    
    in >> word1;    /* read into word1 from beginning */
    std::cout << "tellg " << std::setw(2) << in.tellg() << " - after reading \"" 
                << word1 << "\", rewinding\n";
    
    in.seekg (0);   /* seekg beginning (rewind) */
    std::cout << "tellg " << std::setw(2) << in.tellg() << " - reading into word2\n";
    
    in >> word2;    /* read into word2 from beginning */
    std::cout << "tellg " << std::setw(2) << in.tellg() << " - after reading \"" 
                << word2 << "\", reading final word\n";
    
    in >> word2;    /* read into word2 to end of stringstream, eofbit set */
    std::cout << "tellg " << std::setw(2) << in.tellg() << " - after reading \"" 
                << word2 << "\", eofbit set, tellg() fails, must .clear()\n";
    
    in.clear();     /* clear required before further stringstring operations */
    
    in.seekg (0, std::ios_base::beg);   /* reposition to beginning, 2nd form */
    std::cout << "tellg " << std::setw(2) << in.tellg() << " - rdbuf() -> \"" 
                << in.rdbuf() 
                << "\"\ntellg " << std::setw(2) << in.tellg() << " - reversing: \"";
    
    while (in.tellg() != 0) {   /* playing tell & seek back to beginning */
        in.seekg (-1, std::ios_base::cur);          /* seek -1 from current */
        std::cout << (char)in.get();                /* output character */
        in.seekg (-1, std::ios_base::cur);          /* seek -1 from current */
    }
    std::cout << "\"\ntellg " << std::setw(2) << in.tellg() << " - all done.\n";
}

示例Use/Output

$ ./bin/streambuf_seektellg
string: Hello, world

tellg  0 - reading into word1
tellg  6 - after reading "Hello,", rewinding
tellg  0 - reading into word2
tellg  6 - after reading "Hello,", reading final word
tellg -1 - after reading "world", eofbit set, tellg() fails, must .clear()
tellg  0 - rdbuf() -> "Hello, world"
tellg 12 - reversing "dlrow ,olleH"
tellg  0 - all done.

上面的示例显示了如何处理 stringstring 中 streambuf 位置的操作。从简单的倒带,到使用 goodbit 以外的读取状态处理 streambuf,以及如何使用 tellg()seekg() 的组合从末尾迭代 streambuf 中的每个字符到开始有效地反转输出。


std::stringstream

上使用 seekp() 的其他示例

回应您的评论,除了要使用 tellp()seekp()streambuf 对象必须能够发送输出外,确实没有太大区别给它。这意味着 tellp()seekp() 仅与 std::stringstreamstd::ostringstream 相关,而不与 std::istringstream 相关。因此,为了在上面的示例中定位写入位置,您只需将类型更改为允许输出的两个变体之一。使用 std::stringstream 没问题,例如

    std::stringstream in(str);          /* stringstrem use both seek&tell (gp) */

现在您可以使用 tellp()seekp() 移动 get 并在其中设置写入位置。例如,以下通过再次读取 word1 获取 in 中的当前读取位置,并保存 in.tellg() 报告的位置。然后 seekp() 使用该位置在 "Hello, " 之后定位 "big wide world!" 的写入,例如

    std::cout << "\"\ntellg " << std::setw(2) << in.tellg() 
                << " - seeking with seekp() - add \"big wide world!\"\n";
    
    in >> word1;                            /* move past "Hello," */
    size_t pos = in.tellg();                /* save offset from beginning */
    in.seekp (pos + 1);                     /* seekp() past "Hello, " */
    in << "big wide world!";                /* replace remaining content */
    std::cout << "\nstring: " << in.str() << '\n';  /* output result */

输出变化

更改以上输出的最后一行然后添加新文本将导致:

...
tellg  0 seeking with seekp() - add "big wide world!"

string: Hello, big wide world!

seekg()/tellg()seekp()/tellp()是独立偏移

为了回应您的进一步评论,请务必了解 ...g()...p() 成员函数提供对从 streambuf 对象开头的两个独立偏移量的访问。 seekg()tellg() 与缓冲区中的 读取位置 相关联。 (想想seek...get()tell...get())相比之下,seekp()tellp()与缓冲区中的写入位置相关联。 (想想 seek...put()tell...put()

对该示例的最终输出进行额外更改将使这一点变得清晰。仅将与 /* output results */ 关联的行更改为:

    /* output result */
    std::cout << "\nstring: " << in.str() << "\ntellg  : " << in.tellg()
                << "\ntellp  : " << in.tellp() << '\n';

输出变化

输出现在显示 in.tellg()in.tellp() 在最终写入 in 后报告的独立偏移量,例如

...
string: Hello, big wide world!
tellg  : 6
tellp  : 22

其中in.tellg()报告缓冲区中下一次读取(如in >> word1;)开始的位置,而in.tellp()报告缓冲区中下一次写入的位置将发生,(例如 in << " ... of code";)。添加 in << " ... of code"; 后,从当前 tellg() 位置输出 std::cout << '\n' << in.rdbuf() << '\n'; 将导致:

 big wide world! ... of code

检查一下,如果您还有其他问题,请告诉我。