Find whole word in string, case insensetive.

[IM back!]

im trying find a word in a string but Mach the whole word only ( ex. "pan" should not Mach "expansive" but should Mach "PAN" ). heres what i have so far:

bool ci_equal(const char ch1, const char ch2)
{
    return toupper((unsigned char)ch1) == toupper((unsigned char)ch2);
}

size_t ci_find(const std::string& str1, const std::string& str2)
{
    std::string::const_iterator pos = search(str1. begin ( ), str1. end ( ), str2.begin ( ), str2. end ( ), ci_equal);
    if (pos == str1. end ( ))
    return std::string::npos;
    else
    return pos - str1. begin ( );
}

size_t ci_find_word(const std::string& str1, const std::string& str2)
{
    std::string::const_iterator pos = search(str1. begin ( ), str1. end ( ), str2.begin ( ), str2. end ( ), ci_equal);
    if ( pos != str1.end() && ( pos == str1.begin() || str1.at(pos-str1.begin()-1 ) == ' ') && ( pos == str1.end() || str1.at((pos-str1.begin())+str2.size()-1 ) == ' ') ) return pos - str1. begin ( );
    else return std::string::npos;
}

ci_find works fine but ci_find_word always returns std::string::npos. excuse my ignorance but why do const_iterator, size_ts even exist? couldn't it be done with ints? they are a pain. they have been hell to deal with.

[tabstop]

So have you done any debugging? Call it with something you know will match, and have your function print out your conditions (true/false), the value of pos-str1.begin(), etc.

Also, this:

pos == str1.end()

probably isn't right, since you want to check pos+str2.length().

[EVOEx]

Addition to tabstop (he beat me to it)...

This:

str1.at(pos-str1.begin()-1 ) == ' '

Should probably be

str1.at(pos-1 ) == ' '

[tabstop]

Addition to tabstop (he beat me to it)...

This:

str1.at(pos-str1.begin()-1 ) == ' '

Should probably be

str1.at(pos-1 ) == ' '

I don't think you can throw a const_iterator into at; instead you turn it into a number first (which is what he has done).

[EVOEx]

I don't think you can throw a const_iterator into at; instead you turn it into a number first (which is what he has done).

Ahhh.. pos is an iterator... I missed that. So wouldn't *(pos - 1) work?
I'm not sure if the standard defines the operator- though, on an iterator of a string... So someone should verify.

[tabstop]

For a string, your standard iterators are bidirectional, so - would work. I can't think of a reason (other than paranoia) why *(pos - 1) wouldn't work (it would be a problem if pos was at the beginning of the string, but the short-circuit || should prevent that from happening.

[IM back!]

small debug program:

bool ci_equal(const char ch1, const char ch2)
{
    return toupper((unsigned char)ch1) == toupper((unsigned char)ch2);
}

size_t ci_find(const std::string& str1, const std::string& str2)
{
    std::string::const_iterator pos = search(str1. begin ( ), str1. end ( ), str2.begin ( ), str2. end ( ), ci_equal);
    if (pos == str1. end ( ))
    return std::string::npos;
    else
    return pos - str1. begin ( );
}

size_t ci_find_word(const std::string& str1, const std::string& str2)
{
    std::string::const_iterator pos = search(str1. begin ( ), str1. end ( ), str2.begin ( ), str2. end ( ), ci_equal);
    std::cout<<"pos = "<<pos-str1.begin()<<std::endl;
    if ( pos != str1.end() && ( pos == str1.begin() || str1.at(pos-str1.begin()-1 ) == ' ') && ( pos+str2.length() == str1.end() || str1.at((pos-str1.begin())+str2.size()-1 ) == ' ') ) return pos - str1. begin ( );
    else return std::string::npos;
}

void cd_to_exe_dir( char *argv[] )
{
    std::string path = argv[0];
    int ii = path.length();
    while ( !( path[ii] == '/' || path[ii] == '\\' ) ) ii--;
    path.erase( ii, 100 );
    chdir( path.c_str() );
}

int main( int argc, char *argv[] )
{
    cd_to_exe_dir( argv );
    std::string searchword = "fox";
    std::string searchin   = "this is fierfox 4.0";
    std::string searchin2   = "there is a Fox in here";
    if ( ci_find_word( searchword, searchin ) != std::string::npos ) std::cout<<"found \""<<searchword<<"\" in \""<<searchin<<"\"\n";
    else std::cout<<"did not find \""<<searchword<<"\" in \""<<searchin<<"\"\n";
    if ( ci_find_word( searchword, searchin2 ) != std::string::npos ) std::cout<<"found \""<<searchword<<"\" in \""<<searchin2<<"\"\n";
    else std::cout<<"did not find \""<<searchword<<"\" in \""<<searchin2<<"\"\n";
    return(0);
}

output:

pos = 3
did not find "fox" in "this is fierfox 4.0"
pos = 3
did not find "fox" in "there is a Fox in here"

Press ENTER to continue.

Im still wondering, why where iterators included in the standard?

[tabstop]

Why are iterators included? Because you need things to walk through container objects (and the C++ STL is all about the container object). You need something that's going to walk through a map, and a list, and even a string. What else are you going to use? (The idea is that, for instance, pointers (in an array, for instance) are iterators, in every sense of the term. Iterators extend that concept (but don't necessarily use the same implementation).

You do realize you're searching for "this is fierfox 4.0" inside the string "fox" and not the other way around? (str1 is your container, str2 is what is searched for.)

[IM back!]

Why are iterators included? Because you need things to walk through container objects (and the C++ STL is all about the container object). You need something that's going to walk through a map, and a list, and even a string. What else are you going to use? (The idea is that, for instance, pointers (in an array, for instance) are iterators, in every sense of the term. Iterators extend that concept (but don't necessarily use the same implementation).

why not use ints?

You do realize you're searching for "this is fierfox 4.0" inside the string "fox" and not the other way around? (str1 is your container, str2 is what is searched for.)

thanks for caching that:

still dosent work tho heres the new output:

pos = 12
did not find "fox" in "this is fierfox 4.0"
pos = 11
did not find "fox" in "there is a Fox in here"

Press ENTER to continue.

[tabstop]

why not use ints?

Because ints are nowhere near up to the task? Ints can handle arrays just fine; but anything more complicated (even a plain vanilla linked list, like you would code yourself) and your ints will fold like a cheap suit.

thanks for caching that:

still dosent work tho heres the new output:

pos = 12
did not find "fox" in "this is fierfox 4.0"
pos = 11
did not find "fox" in "there is a Fox in here"

Press ENTER to continue.

And also you need to look after your word for the ending space, not before the last the letter of the word.

[laserlight]

For a string, your standard iterators are bidirectional, so - would work.

A slight correction: the iterators of std::string are random access iterators, for which operator- is defined. Bidirectional iterators do not necessarily provide operator-, thus if you want to be more general you should use std::distance().

[IM back!]

thanks, still no good tho

bool ci_equal(const char ch1, const char ch2)
{
    return toupper((unsigned char)ch1) == toupper((unsigned char)ch2);
}

size_t ci_find(const std::string& str1, const std::string& str2)
{
    std::string::const_iterator pos = search(str1. begin ( ), str1. end ( ), str2.begin ( ), str2. end ( ), ci_equal);
    if (pos == str1. end ( ))
    return std::string::npos;
    else
    return pos - str1. begin ( );
}

size_t ci_find_word(const std::string& str1, const std::string& str2)
{
    std::string::const_iterator pos = search(str1. begin ( ), str1. end ( ), str2.begin ( ), str2. end ( ), ci_equal);
    std::cout<<"pos = "<<pos-str1.begin()<<std::endl;
    if ( pos != str1.end() && ( pos == str1.begin() || str1.at(pos-str1.begin()-1 ) == ' ') && ( pos+str2.length() == str1.end() || str1.at((pos-str1.begin())+str2.size()+1 ) == ' ') ) return pos - str1. begin ( );
    else return std::string::npos;
}

[tabstop]

Checking the letter after where you want the space to be also won't help much, either.