Mapping Words to Line Numbers in Text Files in STL / C++

Following on from the previous post, this example shows an example of how to use an STL multimap to track the line number(s) associated with each word in a text file. This program essentially reads in text line-by-line, while stripping out all occurrences of punctuation and other non-alphanumeric charcters. Each pair is inserted into the multimap container using the insert function. As with the previous posting, which deals with counting the frequency of words in a file, this example also uses the sample Hamlet.txt file. Code listing as follows: [code language="cpp"] #include <iostream> #include <sstream> #include <fstream> #include <map> using namespace std; int main() { const string path = "/home/andy/NetBeansProjects/Hamlet.txt"; //Linux //const string path = "C:\\Dump\\Hamlet.txt"; ifstream input( path.c_str() ); if ( !input ) { cout << "Error opening file." << endl; return 0; } multimap< string, int, less<string> > words; int line; string word; // For each line of text for ( line = 1; input; line++ ) { char buf[ 255 ]; input.getline( buf, 128 ); // Discard all punctuation characters, leaving only words for ( char *p = buf; *p != '\0'; p++ ) { if ( !isalpha( *p ) ) *p = ' '; } istringstream i( buf ); while ( i ) { i >> word; if ( word != "" ) { words.insert( pair<const string,int>( word, line ) ); } } } input.close(); // Output results multimap< string, int, less<string> >::iterator it1; multimap< string, int, less<string> >::iterator it2; for ( it1 = words.begin(); it1 != words.end(); ) { it2 = words.upper_bound( (*it1).first ); cout << (*it1).first << " : "; for ( ; it1 != it2; it1++ ) { cout << (*it1).second << " "; } cout << endl; } return 0; } [/code] Giving the following output. Notice that multiple occurrences of words per line are mapped. Related post: Counting the Number of Words in a Text File in STL / C++