![](/img/trans.png)
[英]pybind11 crashes (segmentation fault (core dumped)) while importing ONNX python module
[英]segmentation fault 11 error while importing csv
我正在做一個函數importcsv(),它接受一個文件名並輸出一個2D數組。 出於某種原因,每當我使用以下版本的importcsv()時,編譯器都會平穩運行,但是可執行文件始終會返回“分段錯誤:11”錯誤。
typedef vector<vector<double> > matrix;
matrix importcsv(string filename)
{
ifstream myfile (filename); //Constructs a stream, and then asssociates the stream with the file "filename"
matrix contents; // Vector which will store the contents of the stream.
int i, j;
while(!myfile.eof())
{
if(myfile.get()==','){++j;}
else if(myfile.get()=='\n'){++i; j=0;}
else{
contents[i][j]=2;}
}
return contents;
}
誰能找到錯誤的根源? 順便說一句,我有以下標題:
#include <fstream>
#include <iostream>
#include <array>
#include <vector>
using namespace std;
由於沒有為contents
分配內存,因此出現“ segmentation fault:11”。 僅當contents
包含contents[i][j]
才有效。
您可以將讀取文件和將矩陣構造為多個部分:
contents
。 這樣,可以簡化程序。 這也可以幫助您輕松地找出問題並加以解決。
typedef vector<vector<double> > matrix;
double readNextNumber(std::istream& str)
{
double n = 0.0;
str >> n;
// Skip until we get ',' or '\n'
while (str)
{
int c = str.getc();
if ( c == ',' || c == '\n' || c == EOF )
break;
}
return n;
}
std::vector<double> importRow(std::ifstram& myfile)
{
std::string line;
std::vector<double> row;
// Read a line as a string.
// Then parse the string using std::istringstream.
// When you have finished parsing the line, you know
// you have finished constructing a row of the matrix.
std::getline(myfile, line);
if ( myfile )
{
std::istringstream str(line);
while (str)
{
double n = readNextNumber(str);
if (str)
{
row.push_back(n);
}
}
}
return row;
}
matrix importcsv(string filename)
{
ifstream myfile (filename); //Constructs a stream, and then asssociates the stream with the file "filename"
matrix contents; // Vector which will store the contents of the stream.
while(!myfile.eof())
{
std::vector<double> row = importRow(myfile);
if (myfile)
{
contents.push_back(row);
}
}
return contents;
}
您尚未定義contents
的大小。 因此,默認情況下,它將是0
元素的vector
。 因此,對operator[]
的調用將導致分段錯誤。
在這里實施其他人的建議時,快速解決方案是在將每個值讀入數組之前使用resize():
//WARNING: THIS PROGRAM UTILIZES C++11
#include <fstream>
#include <iostream>
#include <array>
#include <vector>
#include <cctype>
#include <thread>
using namespace std;
typedef vector<vector<double> > matrix;
matrix importcsv(string filename)
{
ifstream myfile ("wavelengthtorgb.csv"); //Constructs a stream, and then asssociates the stream with the file "filename".
matrix contents {{0.0}};
char nextchar; double data; int i,j;
while(!myfile.eof())
{
myfile.get(nextchar);
if(nextchar==',')
{
++j;
contents[i].resize(j+1);
cout<<"encountered a comma."<<" contents is now " <<i+1<<" x "<<j+1<<'\n';
}
else if(isspace(nextchar))
{
myfile.get(); //You might not need this line - first run with this line, and if there is an error, delete it, and try again.
++i;
contents.resize(i+1);
j=0;
contents[i].resize(j+1);
cout<<"encountered a carriage return."<<" contents is now " <<i+1<<" x "<<j+1<<'\n';
}
else
{
myfile.unget();
myfile >> data;
contents[i][j]=data;
cout<< "encountered a double."<<" contents("<<i<<','<<j<<")="<<data<<'\n';
}
}
return contents;
}
聲明:本站的技術帖子網頁,遵循CC BY-SA 4.0協議,如果您需要轉載,請注明本站網址或者原文地址。任何問題請咨詢:yoyou2525@163.com.