Making a custom istream manipulator - c++

I want to make a custom istream manipulator that reads 2 characters from the input, then skips 2 characters from the input, and does that until it runs out of any input.
For example, if I have code like this:
std::string str;
std::cin >> skipchar >> str;
Where skipchar is my manipulator, if the user enters 1122334455, str should contain 113355.
This is what I've got so far, I don't know what I should put in the while loop condition to make this code work properly:
istream& skipchar(istream& stream)
{
char c;
while(1)
{
for (int i = 0; i < 2; ++i)
stream >> c;
for (int i = 0; i < 2; ++i)
stream.ignore(1, '\0');
}
return stream;
}
Any help would be appreciated.

That's a very nice question. I don't know if it's possible. But I implemented something different which gives you the same short syntax you wanted, by overloading the >> operator with a new class called Skip2. Here is the code (which I really enjoyed writing! :-) )
#include <iostream>
#include <string>
#include <istream>
#include <sstream>
using namespace std;
class Skip2 {
public:
string s;
};
istream &operator>>(istream &s, Skip2 &sk)
{
string str;
s >> str;
// build new string
ostringstream build;
int count = 0;
for (char ch : str) {
// a count "trick" to make skip every other 2 chars concise
if (count < 2) build << ch;
count = (count + 1) % 4;
}
// assign the built string to the var of the >> operator
sk.s = build.str();
// and of course, return this istream
return s;
}
int main()
{
istringstream s("1122334455");
Skip2 skip;
s >> skip;
cout << skip.s << endl;
return 0;
}

It's tricky; istream manipulators don't operate as "filters" on the stream but rather as single-shot operations. The istream manipulators provided by the Standard (noskipws, hex etc.) do their work by setting and clearing flags on the stream, so they only expose functionality that is already available.
However, it is possible to create a filtering streambuf wrapping the streambuf of cin (or any input stream) and use a manipulator to install or remove it:
struct skipbuf : std::streambuf {
std::unique_ptr<std::streambuf> src;
int i;
char buf[4];
skipbuf(std::streambuf* src) : std::streambuf{*src}, src{src} {
setg(buf, buf + 2, buf + 2);
}
std::streambuf* unwrap() {
while (buf + i != gptr())
src->sputbackc(buf[--i]);
return src.release();
}
std::streambuf::int_type underflow() override {
setg(buf, buf, buf + std::min(i = src->sgetn(buf, 4), 2));
return i ? buf[0] : traits_type::eof();
}
};
std::istream& skipchar(std::istream& is) {
is.rdbuf(new skipbuf{is.rdbuf()});
return is;
}
std::istream& noskipchar(std::istream& is) {
if (auto* buf = dynamic_cast<skipbuf*>(is.rdbuf()))
delete (is.rdbuf(buf->unwrap()), buf);
return is;
}
Example of usage:
int main() {
std::istringstream iss{"1122334455 hello"};
std::string s1, s2;
iss >> skipchar >> s1 >> noskipchar >> s2;
std::cout << s1 << ' ' << s2 << std::endl;
}
Expected output (run it online):
113355 hello

Related

read .txt file with doubles and strings - c++ [duplicate]

I am trying to read a file of the following format
id1 1 2 3
id2 2 4 6
id3 5 6 7
...
using this code
Dataset::Dataset(ifstream &file) {
string token;
int i = 0;
while (!file.eof() && (file >> token)){
// read line tokens one-by-one
string ID = token;
vector<int> coords;
while ((file.peek()!='\n') && (!file.eof()) && (file >> token)) {
coords.push_back(atoi(token.c_str()));
}
points.push_back(new Point(ID, coords));
i++;
}
cout << "Loaded " << i << " points." << endl;
}
But it tells me I have read 0 points. What am I doing wrong?
Edit: I am openning this using input_stream.open(input_file) and file.good() returns true.
Edit #2: actually .good() returns true the first time and then false. What is that all about?
Edit #3: GUYS. IT'S FREAKING WINDOWS. When i put the path as Dataset/test.txt by cin it works and when I do it like Dataset\test.txt by the commandline it doesn't...
Now the problem is it seems not stop at new lines!
Edit #4: Freaking windows again! It was peeking '\r' instead of '\n'.
Here's an idea: overload operator>>:
struct Point
{
int x, y, z;
friend std::istream& operator>>(std::istream& input, Point& p);
};
std::istream& operator>>(std::istream& input, Point& p)
{
input >> p.x;
input >> p.y;
input >> p.z;
input.ignore(10000, '\n'); // eat chars until end of line.
return input;
}
struct Point_With_ID
: public Point
{
std::string id;
friend std::istream& operator>>(std::istream& input, Point_With_ID& p);
};
std::istream& operator>>(std::istream& input, Point_With_ID& p)
{
input >> p.id;
input >> static_cast<Point&>(p); // Read in the parent items.
return input;
}
Your input could look like this:
std::vector<Point_With_ID> database;
Point_With_ID p;
while (file >> p)
{
database.push_back(p);
}
I separated the Point class so that it can be used in other programs or assignments.
I managed to make it work by accounting for both '\r' and '\n' endings and ignoring trailing whitespace like this:
Dataset::Dataset(ifstream &file) {
string token;
int i = 0;
while (file >> token){
// read line tokens one-by-one
string ID = token;
vector<int> coords;
while ((file.peek()!='\n' && file.peek()!='\r') && (file >> token)) { // '\r' for windows, '\n' for unix
coords.push_back(atoi(token.c_str()));
if (file.peek() == '\t' || file.peek() == ' ') { // ignore these
file.ignore(1);
}
}
Point p(ID, coords);
points.emplace_back(p);
i++;
// ignore anything until '\n'
file.ignore(32, '\n');
}
cout << "Loaded " << i << " points." << endl;
}
Probably not the best of the solutions suggested but it's working!
You should not use eof() in a loop condition. See Why is iostream::eof inside a loop condition considered wrong? for details. You can instead use the following program to read into the vector of Point*.
#include <iostream>
#include <sstream>
#include <fstream>
#include <vector>
class Point
{
public:
std::string ID = 0;
std::vector<int> coords;
Point(std::string id, std::vector<int> coord): ID(id), coords(coord)
{
}
};
int main()
{
std::vector<Point*> points;
std::ifstream file("input.txt");
std::string line;
int var = 0;
while (std::getline(file, line, '\n'))//read line by line
{
int j = 0;
std::istringstream ss(line);
std::string ID;
ss >> ID;
std::vector<int> coords(3);//create vector of size 3 since we already know only 3 elements needed
while (ss >> var) {
coords.at(j) = var;
++j;
}
points.push_back(new Point(ID, coords));
}
std::cout<<points.size()<<std::endl;
//...also don't forget to free the memory using `delete` or use smart pointer instead
return 0;
}
The output of the above program can be seen here.
Note that if you're using new then you must use delete to free the memory that you've allocated. This was not done in the above program that i have given since i only wanted to show how you can read the data in your desired manner.
You've baked everything up in a complex deserializing constructor. This makes the code hard to understand and maintain.
You have a coordinate, so make class for that, we can call it Coord, that is capable of doing its own deserializing.
You have a Point, which consists of an ID and a coordinate, so make a class for that, that is capable of doing its own deserializing.
The Dataset will then just use the deserializing functions of the Point.
Don't limit deserializing to ifstreams. Make it work with any istream.
Deserializing is often done by overloading operator>> and operator<< for the types involved. Here's one way of splitting the problem up in smaller parts that are easier to understand:
struct Coord {
std::vector<int> data;
// read one Coord
friend std::istream& operator>>(std::istream& is, Coord& c) {
if(std::string line; std::getline(is, line)) { // read until end of line
c.data.clear();
std::istringstream iss(line); // put it in an istringstream
// ... and extract the values:
for(int tmp; iss >> tmp;) c.data.push_back(tmp);
}
return is;
}
// write one Coord
friend std::ostream& operator<<(std::ostream& os, const Coord& c) {
if(not c.data.empty()) {
auto it = c.data.begin();
os << *it;
for(++it; it != c.data.end(); ++it) os << ' ' << *it;
}
return os;
}
};
struct Point {
std::string ID;
Coord coord;
// read one Point
friend std::istream& operator>>(std::istream& is, Point& p) {
return is >> p.ID >> p.coord;
}
// write one Point
friend std::ostream& operator<<(std::ostream& os, const Point& p) {
return os << p.ID << ' ' << p.coord;
}
};
struct Dataset {
std::vector<Point> points;
// read one Dataset
friend std::istream& operator>>(std::istream& is, Dataset& ds) {
ds.points.clear();
for(Point tmp; is >> tmp;) ds.points.push_back(std::move(tmp));
if(!ds.points.empty()) is.clear();
return is;
}
// write one Dataset
friend std::ostream& operator<<(std::ostream& os, const Dataset& ds) {
for(auto& p : ds.points) os << p << '\n';
return os;
}
};
If you really want a deserializing constructor in Dataset you just need to add these:
Dataset() = default;
Dataset(std::istream& is) {
if(!(is >> *this))
throw std::runtime_error("Failed reading Dataset");
}
You can then open your file and use operator>> to fill the Dataset and operator<< to print the Dataset on screen - or to another file if you wish.
int main() {
if(std::ifstream file("datafile.dat"); file) {
if(Dataset ds; file >> ds) { // populate the Dataset
std::cout << ds; // print the result to screen
}
}
}
Demo

How do I include an int inside a getline();

Just a simple question, how do I include int values inside a getline()? I have searched online but couldn't find any that helps me. I am reading off a txt file. It is a row of numbers. For eg: 1,2,3,4,5. I am hoping that I can apply these int values anywhere so that the only way I can change the values is through the txt file.
I decided to use getline() but realise that I cannot use an integer. I am sorry, I am new to this C++. I hope that you can tell me where I went wrong.
Thanks!
This is my struct:
struct vacancyData {
int CCSpot;
int SNSpot;
int TPSpot;
int SCSpot;
int DRSpot;
};
This is my code:
ifstream File2;
File2.open("Vacancy.txt");
vector<vacancyData> v1;
vacancyData f;
while (getline(File2, f.CCSpot, ','))
{
getline(File2, f.SNSpot, ',');
getline(File2, f.TPSpot, ',');
getline(File2, f.SCSpot, ',');
getline(File2, f.DRSpot, '\n');
v1.push_back(f);
}
The direct answer to the question
How do I include an int inside a getline()
is: This is not possible at all.
std::getline is basically used to read a std::string from a stream, until a delimiter is found. In most cases, and that is also a default argument, the delimiter is '\n'. And with that, a complete line is read into a std::string. Please read here for a description.
If your input data is OK in most cases, then no std::getline is needed. Basic input validation can also be done directly. Please see the below example code:
#include <iostream>
#include <fstream>
#include <string>
#include <vector>
struct VacancyData {
int CCSpot;
int SNSpot;
int TPSpot;
int SCSpot;
int DRSpot;
};
int main() {
// Filename
const std::string fileName{ "r:\\vacancy.txt" };
// Open file and check, if it could be opened. Use C++17 if-statement with initializer
if (std::ifstream vacancyFileStream{ fileName }; vacancyFileStream) {
// Here we will stor our data
std::vector<VacancyData> all{};
// Temps, to check, if delimiter is comma
char c1{}, c2{}, c3{}, c4{};
// This is one for loop and will read the complet file and do basic input validation
for (VacancyData vc{};
(vacancyFileStream >> vc.CCSpot >> c1 >> vc.SNSpot >> c2 >> vc.TPSpot >> c3 >> vc.SCSpot >> c4 >> vc.DRSpot) &&
c1 == ',' && c2 == ',' && c3 == ',' && c4 == ',';
all.push_back(vc))
; // Empt loop body. No statement within for loop body
// Ws there an error and could all data be read?
if (vacancyFileStream.fail() || not vacancyFileStream.eof())
std::cerr << "\n\nErorw hile reading input data\n\n";
for (const VacancyData& vc : all)
std::cout << vc.CCSpot << '\t' << vc.SNSpot << '\t' << vc.TPSpot << '\t' << vc.SCSpot << '\t' << vc.DRSpot << '\n';
}
else std::cerr << "\n\nEror: cannot open source file '" << fileName << "'\n\n";
return 0;
}
But, it is considere good practice to first read a complete line, then split it up in parts and then convert it to your structure elements.
Additionally: In C++ we often use an object oriented approach. Meaning, and objects methods, operating with the data are encapsulated in a class/struct.
In your case, we would overwrite the extractor >> and inserter operator <<.
This would then look like that (I am using C++17):
#include <iostream>
#include <fstream>
#include <string>
#include <vector>
#include <iterator>
#include <sstream>
struct VacancyData {
int CCSpot;
int SNSpot;
int TPSpot;
int SCSpot;
int DRSpot;
friend std::istream& operator >> (std::istream& is, VacancyData& vc) {
char c{};
return is >> vc.CCSpot >> c >> vc.SNSpot >> c >> vc.TPSpot >> c >> vc.SCSpot >> c >> vc.DRSpot;
}
friend std::ostream& operator << (std::ostream& os, const VacancyData& vc) {
return os << vc.CCSpot << '\t' << vc.SNSpot << '\t' << vc.TPSpot << '\t' << vc.SCSpot << '\t' << vc.DRSpot;
}
};
int main() {
// Filename
const std::string fileName{ "r:\\vacancy.txt" };
// Open file and check, if it could be opened. Use C++17 if-statement with initializer
if (std::ifstream vacancyFileStream{ fileName }; vacancyFileStream) {
// Here we will store our data. Read complete file and parse it
std::vector all(std::istream_iterator< VacancyData>(vacancyFileStream), {});
// Ws there an error and could all data be read?
if (vacancyFileStream.fail() || not vacancyFileStream.eof())
std::cerr << "\n\nErorw hile reading input data\n\n";
// Show output
for (const VacancyData& vc : all)
std::cout << vc << '\n';
}
else std::cerr << "\n\nEror: cannot open source file '" << fileName << "'\n\n";
return 0;
}
Now, if you want to change the reading of your data, you can use any method that you want. For example, you can read a complete line, using std::getline use any method shown below to split the input string into parts. Then you can use any method to convert that string parts into an integer for your struct. You will just change the body of your extractor operator. Nothing else will be affected. That sis the beauty of encapsulation.
Regarding: Splitting a string into tokens is a very old task. There are many many solutions available. All have different properties. Some are difficult to understand, some are hard to develop, some are more complex, slower or faster or more flexible or not.
Alternatives
Handcrafted, many variants, using pointers or iterators, maybe hard to develop and error prone.
Using old style std::strtok function. Maybe unsafe. Maybe should not be used any longer
std::getline. Most used implementation. But actually a "misuse" and not so flexible
Using dedicated modern function, specifically developed for this purpose, most flexible and good fitting into the STL environment and algortithm landscape. But slower.
Please see 4 examples in one piece of code.
#include <iostream>
#include <fstream>
#include <sstream>
#include <string>
#include <regex>
#include <algorithm>
#include <iterator>
#include <cstring>
#include <forward_list>
#include <deque>
using Container = std::vector<std::string>;
std::regex delimiter{ "," };
int main() {
// Some function to print the contents of an STL container
auto print = [](const auto& container) -> void { std::copy(container.begin(), container.end(),
std::ostream_iterator<std::decay<decltype(*container.begin())>::type>(std::cout, " ")); std::cout << '\n'; };
// Example 1: Handcrafted -------------------------------------------------------------------------
{
// Our string that we want to split
std::string stringToSplit{ "aaa,bbb,ccc,ddd" };
Container c{};
// Search for comma, then take the part and add to the result
for (size_t i{ 0U }, startpos{ 0U }; i <= stringToSplit.size(); ++i) {
// So, if there is a comma or the end of the string
if ((stringToSplit[i] == ',') || (i == (stringToSplit.size()))) {
// Copy substring
c.push_back(stringToSplit.substr(startpos, i - startpos));
startpos = i + 1;
}
}
print(c);
}
// Example 2: Using very old strtok function ----------------------------------------------------------
{
// Our string that we want to split
std::string stringToSplit{ "aaa,bbb,ccc,ddd" };
Container c{};
// Split string into parts in a simple for loop
#pragma warning(suppress : 4996)
for (char* token = std::strtok(const_cast<char*>(stringToSplit.data()), ","); token != nullptr; token = std::strtok(nullptr, ",")) {
c.push_back(token);
}
print(c);
}
// Example 3: Very often used std::getline with additional istringstream ------------------------------------------------
{
// Our string that we want to split
std::string stringToSplit{ "aaa,bbb,ccc,ddd" };
Container c{};
// Put string in an std::istringstream
std::istringstream iss{ stringToSplit };
// Extract string parts in simple for loop
for (std::string part{}; std::getline(iss, part, ','); c.push_back(part))
;
print(c);
}
// Example 4: Most flexible iterator solution ------------------------------------------------
{
// Our string that we want to split
std::string stringToSplit{ "aaa,bbb,ccc,ddd" };
Container c(std::sregex_token_iterator(stringToSplit.begin(), stringToSplit.end(), delimiter, -1), {});
//
// Everything done already with range constructor. No additional code needed.
//
print(c);
// Works also with other containers in the same way
std::forward_list<std::string> c2(std::sregex_token_iterator(stringToSplit.begin(), stringToSplit.end(), delimiter, -1), {});
print(c2);
// And works with algorithms
std::deque<std::string> c3{};
std::copy(std::sregex_token_iterator(stringToSplit.begin(), stringToSplit.end(), delimiter, -1), {}, std::back_inserter(c3));
print(c3);
}
return 0;
}
In modern C++ you would probably do:
#include <iostream>
#include <fstream>
#include <string>
#include <vector>
#include <iterator>
#include <sstream>
#include <regex>
// Regex for integer
const std::regex re{R"(([-+]?\d+))"};
struct VacancyData {
int CCSpot;
int SNSpot;
int TPSpot;
int SCSpot;
int DRSpot;
friend std::istream& operator >> (std::istream& is, VacancyData& vc) {
// Read a complete line and check, if that worked
if (std::string line{}; std::getline(is, line)) {
// Split the string into parts. The parts will definitely contain a integer
std::vector part(std::sregex_token_iterator(line.begin(), line.end(), re), {});
// Sanity check. Could we read 5 values?
if (not (part.size() == 5u)) {
std::cerr << "\n\nError while parsing line '" << line << '\n';
is.setstate(std::ios::failbit);
}
else {
// Splitting the string worked. We will have intergers in the parts. stoi will not fail
vc.CCSpot = std::stoi(part[0]);
vc.SNSpot = std::stoi(part[1]);
vc.TPSpot = std::stoi(part[2]);
vc.SCSpot = std::stoi(part[3]);
vc.DRSpot = std::stoi(part[4]);
}
}
return is;
}
friend std::ostream& operator << (std::ostream& os, const VacancyData& vc) {
return os << vc.CCSpot << '\t' << vc.SNSpot << '\t' << vc.TPSpot << '\t' << vc.SCSpot << '\t' << vc.DRSpot;
}
};
int main() {
// Filename
const std::string fileName{ "r:\\vacancy.txt" };
// Open file and check, if it could be opened. Use C++17 if-statement with initializer
if (std::ifstream vacancyFileStream{ fileName }; vacancyFileStream) {
// Here we will store our data. Read complete file and parse it
std::vector all(std::istream_iterator< VacancyData>(vacancyFileStream), {});
// Show output
for (const VacancyData& vc : all)
std::cout << vc << '\n';
}
else std::cerr << "\n\nEror: cannot open source file '" << fileName << "'\n\n";
return 0;
}
But, there are tons of different possible solutions. And everybody can select whatever.
All above needs to be compiles with C++17.
You can use std::getline to do this, but you need use a std::string and then to convert it to an integer. One way is to use the std::stoi function.
Example:
#include <algorithm>
#include <fstream>
#include <iostream>
#include <iterator>
#include <string>
#include <vector>
struct vacancyData {
int CCSpot;
int SNSpot;
int TPSpot;
int SCSpot;
int DRSpot;
};
// Overload operator>> for reading a "vacancyData" struct from an istream (like a file)
std::istream& operator>>(std::istream& is, vacancyData& vd) {
std::string tmp;
try {
if(std::getline(is, tmp, ',')) {
vd.CCSpot = std::stoi(tmp);
if(std::getline(is, tmp, ',')) {
vd.SNSpot = std::stoi(tmp);
if(std::getline(is, tmp, ',')) {
vd.TPSpot = std::stoi(tmp);
if(std::getline(is, tmp, ',')) {
vd.SNSpot = std::stoi(tmp);
if(std::getline(is, tmp, ',')) {
vd.SCSpot = std::stoi(tmp);
if(std::getline(is, tmp)) {
vd.DRSpot = std::stoi(tmp);
}
}
}
}
}
}
}
catch(...) { // one of the stoi calls failed
is.setstate(std::ios::failbit);
}
return is;
}
int main() {
std::ifstream File2("Vacancy.txt");
if(File2) {
// construct the vector using iterators:
std::vector<vacancyData> v1(std::istream_iterator<vacancyData>(File2),
std::istream_iterator<vacancyData>{});
// use the filled vector "v1" ...
}
}
But since there is built-in support for extracting ints directly from istreams, I suggest using that instead.
Example:
// A small support class to "eat" separators, like comma and newline
struct eater { char ch; };
std::istream& operator>>(std::istream& is, eater& e) {
if(is.peek() == e.ch) is.ignore(); // if the next char is the expected, skip it
else is.setstate(std::ios::failbit); // else set the failbit
return is;
}
std::istream& operator>>(std::istream& is, vacancyData& vd) {
eater comma{','};
eater newline{'\n'};
return is >>
vd.CCSpot >> comma >>
vd.SNSpot >> comma >>
vd.TPSpot >> comma >>
vd.SCSpot >> comma >>
vd.DRSpot >> newline;
}

how can I input a literal string into istream? [duplicate]

How to achieve scanf("%d # %d",&a,&b);sort of effect with cin in C++ ?
You can skip the # by extracting it into a character:
std::istringstream iss("10 # 20");
int main()
{
int a, b; char hash;
iss >> a >> hash >> b;
assert(a == 10 && b == 20);
}
You could create your own stream manipulator. It is fairly easy.
#include <ios>
#include <iostream>
using namespace std;
// skips the number of characters equal to the length of given text
// does not check whether the skipped characters are the same as it
struct skip
{
const char * text;
skip(const char * text) : text(text) {}
};
std::istream & operator >> (std::istream & stream, const skip & x)
{
ios_base::fmtflags f = stream.flags();
stream >> noskipws;
char c;
const char * text = x.text;
while (stream && *text++)
stream >> c;
stream.flags(f);
return stream;
}
int main()
{
int a, b;
cin >> a >> skip(" # ") >> b;
cout << a << ", " << b << endl;
return 0;
}
There isn't a direct function inside the istream class that mimics it, unfortunately. There are functions that you might be able to use to manipulate the stream and get the correct input, but I'm not familiar with how they work so I couldn't tell you how.
My best suggestion on how I would personally do it is to use getline() to put the input into a string and then from there I would do a few checks to see if it matches the format. So in your case I would grab the first substring up until the first space, make sure it's a valid decimal, check to make sure the pound sign ('#') is in the correct spot, then grab the ending number to make sure it's valid. If any one of those three objects are incorrect I would set some boolean variable to false to kick out or return or something to indicate that the input was invalid and not the correct format.
Pseudo Code:
...
getline(cin,myStr);
while(!formatCheck(myStr))
{
cout<<"Not valid format for input";
getline(cin,myStr);
}
...
bool formatCheck(string str)
{
string firstPart=str.subString(0,firstSpaceLocation);
string middle=str[firstSpaceLocation+1];
string lastPart=str.subString(firstSpaceLocation+3,end);
if(first part not a valid number || middle!="#" || last part not a valid number)
{
return false;
}
return true;
}
Here's another way. You can classify # as a whitespace character through the std::ctype<char> facet imbued in the locale:
#include <iostream>
#include <sstream>
#include <vector>
namespace detail
{
enum options { add, remove };
class ctype : public std::ctype<char>
{
private:
static mask* get_table(const std::string& ws, options opt)
{
static std::vector<mask> table(classic_table(),
classic_table() + table_size);
for (char c : ws)
{
if (opt == add)
table[c] |= space;
else if (opt == remove)
table[c] &= ~space;
}
return &table[0];
}
public:
ctype(const std::string& ws, options opt)
: std::ctype<char>(get_table(ws, opt)) { }
};
}
class adjustws_impl
{
public:
adjustws_impl(const std::string& ws, detail::options opt) :
m_ws(ws),
m_opt(opt)
{ }
friend std::istream& operator>>(std::istream& is,
const adjustws_impl& manip)
{
is.imbue(std::locale(is.getloc(),
new detail::ctype(manip.m_ws, manip.m_opt)));
return is;
}
private:
std::string m_ws;
detail::options m_opt;
};
adjustws_impl setws(const std::string& ws)
{
return adjustws_impl(ws, detail::add);
}
adjustws_impl unsetws(const std::string& ws)
{
return adjustws_impl(ws, detail::remove);
}
int main()
{
std::istringstream iss("10 # 20");
int a, b;
iss >> setws("#");
iss >> a >> b;
iss >> unsetws("#");
std::cout << a << ' ' << b; // 10 20
}
You can skip the #, or any single character, by using std::istream::ignore
std::istringstream sstr("1024 # 768");
int main()
{
int a, b;
sstr >> a;
sstr.ignore(256,'#'); // ignore until hash character
sstr >> b;
std::cout << "a: " << a << " b: " << b << std::endl;
}

Overload ifstream in c++

I am trying to read complex numbers of the form x + y*i from a file which looks like this:
2 + 3i
4 + 5i
If I implement it like this it only works for the first line and I would like to be able to read again in the same manner the second line. Any ideas?
istream& operator>>(istream& in, Complex& c) {
in >> c.r >> c.i;
return in;
}
EDIT: Do not throw exceptions directly, as this is not the usual way of doing thing with iostreams.
EDIT: Process the sign character separately so that spaces in the input are allowed.
A quick and dirty solution:
istream& operator>>(istream& in, Complex& c)
{
char sign;
char suffix;
in >> c.r >> sign >> c.i >> suffix;
if ((sign != '+' && sign != '-') || suffix != 'i') {
in.setstate(ios::failbit);
}
if (sign == '-') {
c.i = -c.i;
}
return in;
}
You need to make sure that you read both the "+" and the "i" when processing the input.
The following implementation works:
struct complex
{
double r;
double i;
};
std::istream& operator>>(std::istream& in, complex& c)
{
char plus, i;
in >> c.r >> plus >> c.i >> i;
return in;
}
std::string initdata = "2 + 3i\n4 + 5i\n";
int main()
{
std::stringstream ss(initdata);
std::vector<complex> values;
std::istream_iterator<complex> begin(ss), end;
std::copy(begin, end, std::back_inserter<std::vector<complex>>(values));
}
I would do it like this :
#include <iostream>
struct complex
{
int r, i;
};
int main ()
{
complex co;
char c;
while ((std::cin >> co.r >> c /* '+' */ >> co.i >> c /* 'i' */))
{
std::cout << co.r << ' ' << co.i << "i\n";
}
}
This should work for you:
#include <iostream>
#include <sstream>
int main(int argc, char *argv[])
{
std::istringstream s(
"2 + 3i\n"
"4 + 5i\n"
);
char sgn;
double r;
double i;
while(s >> r >> sgn >> i) {
if(s.get() != 'i') {
std::cerr << "Missing i\n";
return -1;
}
std::cout << r << sgn << i << "i\n";
}
return 0;
}
Note: The space before the imaginary part and the trailing i are breaking the input.

Skipping expected characters like scanf() with cin

How to achieve scanf("%d # %d",&a,&b);sort of effect with cin in C++ ?
You can skip the # by extracting it into a character:
std::istringstream iss("10 # 20");
int main()
{
int a, b; char hash;
iss >> a >> hash >> b;
assert(a == 10 && b == 20);
}
You could create your own stream manipulator. It is fairly easy.
#include <ios>
#include <iostream>
using namespace std;
// skips the number of characters equal to the length of given text
// does not check whether the skipped characters are the same as it
struct skip
{
const char * text;
skip(const char * text) : text(text) {}
};
std::istream & operator >> (std::istream & stream, const skip & x)
{
ios_base::fmtflags f = stream.flags();
stream >> noskipws;
char c;
const char * text = x.text;
while (stream && *text++)
stream >> c;
stream.flags(f);
return stream;
}
int main()
{
int a, b;
cin >> a >> skip(" # ") >> b;
cout << a << ", " << b << endl;
return 0;
}
There isn't a direct function inside the istream class that mimics it, unfortunately. There are functions that you might be able to use to manipulate the stream and get the correct input, but I'm not familiar with how they work so I couldn't tell you how.
My best suggestion on how I would personally do it is to use getline() to put the input into a string and then from there I would do a few checks to see if it matches the format. So in your case I would grab the first substring up until the first space, make sure it's a valid decimal, check to make sure the pound sign ('#') is in the correct spot, then grab the ending number to make sure it's valid. If any one of those three objects are incorrect I would set some boolean variable to false to kick out or return or something to indicate that the input was invalid and not the correct format.
Pseudo Code:
...
getline(cin,myStr);
while(!formatCheck(myStr))
{
cout<<"Not valid format for input";
getline(cin,myStr);
}
...
bool formatCheck(string str)
{
string firstPart=str.subString(0,firstSpaceLocation);
string middle=str[firstSpaceLocation+1];
string lastPart=str.subString(firstSpaceLocation+3,end);
if(first part not a valid number || middle!="#" || last part not a valid number)
{
return false;
}
return true;
}
Here's another way. You can classify # as a whitespace character through the std::ctype<char> facet imbued in the locale:
#include <iostream>
#include <sstream>
#include <vector>
namespace detail
{
enum options { add, remove };
class ctype : public std::ctype<char>
{
private:
static mask* get_table(const std::string& ws, options opt)
{
static std::vector<mask> table(classic_table(),
classic_table() + table_size);
for (char c : ws)
{
if (opt == add)
table[c] |= space;
else if (opt == remove)
table[c] &= ~space;
}
return &table[0];
}
public:
ctype(const std::string& ws, options opt)
: std::ctype<char>(get_table(ws, opt)) { }
};
}
class adjustws_impl
{
public:
adjustws_impl(const std::string& ws, detail::options opt) :
m_ws(ws),
m_opt(opt)
{ }
friend std::istream& operator>>(std::istream& is,
const adjustws_impl& manip)
{
is.imbue(std::locale(is.getloc(),
new detail::ctype(manip.m_ws, manip.m_opt)));
return is;
}
private:
std::string m_ws;
detail::options m_opt;
};
adjustws_impl setws(const std::string& ws)
{
return adjustws_impl(ws, detail::add);
}
adjustws_impl unsetws(const std::string& ws)
{
return adjustws_impl(ws, detail::remove);
}
int main()
{
std::istringstream iss("10 # 20");
int a, b;
iss >> setws("#");
iss >> a >> b;
iss >> unsetws("#");
std::cout << a << ' ' << b; // 10 20
}
You can skip the #, or any single character, by using std::istream::ignore
std::istringstream sstr("1024 # 768");
int main()
{
int a, b;
sstr >> a;
sstr.ignore(256,'#'); // ignore until hash character
sstr >> b;
std::cout << "a: " << a << " b: " << b << std::endl;
}