Best way to find a whitespace-delimited word in a

问题:

example: "select * from somewhere where x = 1"

I want to find the whitespace-delimited "where", but not the "where" within "somewhere". In the example "where" is delimited by spaces, but it could be carriage returns, tabs etc.

Note: I know regex would make it easy to do (the regex equivalent would be "\bwhere\b"), but I don't want to add a regex library to my project just to do this.

回答1:

If you wanted to use the pure MFC method of string manipulation, then this should work:

CString strSql = _T("select * from somewhere where x = 1");

int nTokenPos = 0;
CString strToken = strSql.Tokenize(_T(" \r\n\t"), nTokenPos);

while (!strToken.IsEmpty())
{
    if (strToken.Trim().CompareNoCase(_T("where")) == 0)
        return TRUE; // found
    strToken = strSql.Tokenize(_T(" \r\n\t"), nTokenPos);
}

return FALSE; // not found

回答2:

I don't know the first thing about MFC and CString, but in plain C++ I'd stuff that string into an std::istringstream and read from that into a string:

std::istringstream iss("select * from somewhere where x = 1");
std::string word;
do {
  iss >> word;
  if( !iss ) throw "blah blah!";
} while( word != "where" );

I suppose CString does have stream operators overloaded with the appropriate semantics, but, as I said, I don't know it.

回答3:

Pretty fast and doesn't need to construct heavy istringstream objects.

CString str = L"select * from somewhere where x = 1";
CString to_find = L"where";
int i =0;
while ( i < str.GetLength() ) {
    i = str.Find(to_find, i);
    if ( i == -1 ) break;
    if ( (i == 0 || !isalpha(str[i-1])) && (i == str.GetLength()-to_find.GetLength() || !isalpha(str[i+to_find.GetLength()])) ) break;
    i+=to_find.GetLength();
}
if ( i >= 0 && i < str.GetLength() ) // found
else // not found

回答4:

If you are using Visual Studio 2008 with Feature Pack you already have std::tr1::regex in C++.

回答5:

Well, regex is the correct answer. If you don't want to include regex, you will have to try and implement a bit of regex yourself and search for " where ".

Your only other option I guess would be to search for "where", then check the character before and after the match to see if they are whitespace characters.