Removing Puctuation from Strings

[swanley007]

Hello, I am having som problems with a program I am writing.

I am to remove punctuation from words read in from a file.

"-" are ok

EX.

infile >> temp_word;

for(i=0; i<strlen(temp_word.c_str());i++)
{
      if(isalpha(temp_word[i])
      {
              temp_word[i] = temp_word[i];
      }
}

The strings print out with

 cout << temp_word;

but they have punctuation in them still


PLEASE HELP


Also I wouls like to use STL's like

 isalpha()

if poss.

[hk_mp5kpdw]

Look up the remove_if STL function (in the <algorithm> header) and the ispunct function.

[swanley007]

I tried using an

(!ispucnct)

but it didn't do anything

[Daved]

First, the string class has a length() method, there is no need to use strlen, just call temp_word.length().

Second, your problem is that you are assigning the character to itself in the same string. You need two variables, the source and the target. If the character is not punctuation (!ispunct(temp_word[i])), then add it to the target string (using += would be easiest). Otherwise don't add it. The resulting string will have the punctuation removed.

[swanley007]

I found the answer on google:

size_t pos; // Position holder

while( (pos = s.find_first_of("`~!@#$%^&*()-_=+{}|[]\\:;\"\'<>,.?/\t\r\n")) != string::npos)
s.erase(pos, 1);

where s = name of the string

[rockytriton]

looks very inefficient to me, since it keeps looping through the whole string until all punctuation is removed. I'd say do it all in a single pass with a for loop and use a stringstream for output or something.. something like this....

std::stringstream buff;

for (int i=0; i<s.size(); i++)
{
    if (isalpha(s[i]))
        buff << s[i];
}

return buff.str();

[byte]

Have you tried using ascii's integer equivalent? Using if statement, test whether a particular element belongs to a certain range of punctuation symbols (ascii).

[laserlight]

I think the solution that hk_mp5kpdw had in mind would be to define:

bool isPunctChar(char ch) {
	return std::ispunct(ch);
}

Then for a std::string str use:

str.erase(std::remove_if(str.begin(), str.end(), isPunctChar), str.end());

After including the <algorithm>, <cctype> and <string> standard headers.

std::ispunct() takes an int, so we define isPunctChar() to take a char.

std::remove_if() operates on a range and removes the characters in str for which isPunctChar() returns true. Okay, it doesnt quite remove them, but now you have a substring of characters that dont contain the unwanted characters, and a substring of characters that may be discarded. std::remove_if() returns an iterator to this range of unwanted characters, so str.erase() is used to, er, erase them.

Of course, your (original) requirements dont exactly match std::ispunct() since you allow dashes, but then isPunctChar() can be modified to allow them.