Char * to int

[Kay]

Is it possible to convert char * to integer ?
if yes, how to do it ?

 

[Phlip]

Is it possible to convert char * to integer ?
if yes, how to do it ?

char const * z = "12";

stringstream stream = z;
int q = 0;

if (z >> q)
cout << "yes" << endl;

(strtol() also works.)

 

[Mike Wahler]

Is it possible to convert char * to integer ?

No, not portably.

if yes, how to do it ?

Perhaps you want to convert the textual representation
of a number to an integer type. If so, use a stringstream.

#include <iostream>
#include <sstream>

int main()
{
const char *x = "123";
std::istringstream iss(x);
int i(0);

if(iss >> i)
std::cout << i << '\n';
else
std::cerr << "unable to convert\n";

return 0;
}


-Mike

 

[David Hilsee]

Is it possible to convert char * to integer ?
if yes, how to do it ?

The technique listed in the FAQ for converting std::strings to numbers will
also work for C-style strings. See the FAQ
(http://www.parashift.com/c++-faq-lite/), section 38 ("Miscellaneous
technical issues"), question 2 ("How do I convert a std::string to a
number?").

 

[Method Man]

char const * z = "12";

stringstream stream = z;
int q = 0;

if (z >> q)
cout << "yes" << endl;

(strtol() also works.)

Isn't using 'atoi' function in <stdlib.h> much simpler?

 

[Rolf Magnus]

Isn't using 'atoi' function in <stdlib.h> much simpler?

Yes, but it's also much worse, because it has a design bug. It doesn't
provide any way to find out whether the conversion was successful.

 

[Mike Wahler]

Isn't using 'atoi' function in <stdlib.h> much simpler?

Yes, but it cannot be used safely. (Does not protect from
overflow, which would cause undefined behavior. Also, a
valid result of zero cannot be distinguished from an error,
which 'atoi()' indicates by returning zero.

-Mike

 

[Default User]

Isn't using 'atoi' function in <stdlib.h> much simpler?

Simpler, perhaps. It has no error handling capability though. It's
generally recommended to you strtol() if you want to use C-style
functionality.


Brian Rodenborn

 

[Method Man]

Yes, but it cannot be used safely. (Does not protect from
overflow, which would cause undefined behavior. Also, a
valid result of zero cannot be distinguished from an error,
which 'atoi()' indicates by returning zero.

-Mike

Hmm, I would just write my own version of 'atoi' then. At the least, it's
good practice for a C/C++ job interview.

 

[Mike Wahler]

Hmm, I would just write my own version of 'atoi' then. At the least, it's
good practice for a C/C++ job interview.

Why would you not use 'strto(u)l()'?

If a candidate wasted time reinventing what the standard library
already has, I would not be impressed. I'd be more impressed with
one who knows which tools are available, and when/how/where to apply
them.

Anyone familiar with C or C++ will recognize 'strtol()' at a glance,
whereas upon encountering a 'home-grown' conversion function, one
would need to find out what the darn thing actually does. Also, a
standard library function has a far better chance of being fully
debugged and robust.


-Mike

 

[Daniel T.]

Is it possible to convert char * to integer ?
if yes, how to do it ?

<http://www.boost.org/libs/conversion/lexical_cast.htm>

int main(int argc, char * argv[])
{
using boost::lexical_cast;
using boost::bad_lexical_cast;

std::vector<short> args;

while(*++argv)
{
try
{
args.push_back(lexical_cast<short>(*argv));
}
catch(bad_lexical_cast &)
{
args.push_back(0);
}
}
...
}

There you go...

 

[Method Man]

Why would you not use 'strto(u)l()'?

If a candidate wasted time reinventing what the standard library
already has, I would not be impressed. I'd be more impressed with
one who knows which tools are available, and when/how/where to apply
them.

Anyone familiar with C or C++ will recognize 'strtol()' at a glance,
whereas upon encountering a 'home-grown' conversion function, one
would need to find out what the darn thing actually does. Also, a
standard library function has a far better chance of being fully
debugged and robust.


-Mike

Yes, you're right.

I guess I've always used 'atoi'. This is what I was taught back in my 1st
year university CS class for converting any C-style strings to int.

Regarding the interview question. I've actually been asked to specifically
implement 'atoi' in two separate interviews. I suppose the context was
simply that it's a good test for one's pointer knowledge and error handling
ability with strings.

 

[Karl Heinz Buchegger]

I guess I've always used 'atoi'. This is what I was taught back in my 1st
year university CS class for converting any C-style strings to int.

Regarding the interview question. I've actually been asked to specifically
implement 'atoi' in two separate interviews. I suppose the context was
simply that it's a good test for one's pointer knowledge and error handling
ability with strings.

int atoi( char* String )
{
int Result;
sscanf( "%d", Result );
return Result;
}



I guess this is not what your interviewer expected you to do

 

[Mike Wahler]

int atoi( char* String )
{
int Result;
sscanf( "%d", Result );

return (sscanf(String, "%d", &Result) == 1) ? Result : 0;

return Result;
}

-Mike

 

[Phlip]

Hmm, I would just write my own version of 'atoi' then. At the least,
it's good practice for a C/C++ job interview.

If asked to write atoi() at a job interview, I would match its exact ISO
specification, verbally point out all its minor flaws, reveal when and when
not I would use it, and write wrapper code that defends its calls.

(Not sure if I could match the performance with lexical_cast, though!)

 

[Rolf Magnus]

If asked to write atoi() at a job interview, I would match its exact ISO
specification, verbally point out all its minor flaws,

It has minor flaws too?

reveal when and when not I would use it, and write wrapper code that
defends its calls.

If asked to write it, I'd write it. Nothing more. If they want to know more,
they can ask. Well, maybe I'd just note that atoi() is a bad idea because
has a major design bug.
It's just like in oral exams. Don't answer questions that have not been
asked.