Error in output in c++ program - c++

Question

The purpose of the program is to read a phrase from a file into a vector and convert the phrase into Pig Latin. When the translated phrase is outputted in Pig Latin, an additional "ay" is added after the phrase (which is not supposed to happen). Can anyone spot why this is happening? It is important that I fix this because it affects the total letters and total characters of the Pig Latin phrase that I need to output. Also, I'm not asking anyone to write any code for me, but any tips on how to make my code less redundant. A portion of my grade for programs is efficiency, which I usually lose points on.
Here's the code:
#include <iostream>
#include <fstream>
#include <vector>
#include <algorithm>
#include <sstream>
#include <cctype>
#include <stdio.h>
#include <ctype.h>
using namespace std;
int main()
{
ifstream in;
string word, fileName;
vector <string> phrase;
int length = 0, index = 0;
int totalWords = -1, totalLetters = -3, totalChars;
cout << "PIG LATIN PROGRAM" << endl;
cout << "Which file are you accessing? : ";
cin >> fileName;
fileName += ".txt";
in.open(fileName);
if (in.fail()) cout << "\nFile not found!" << endl;
while(getline(in, word)) phrase.push_back(word);
cout << "Original Phrase: " << phrase[0] << endl;
istringstream iss(phrase[0]);
cout << "Pig Latin phrase: ";
do {
string OGword;
string PLword;
for (int i=0; i < phrase.size(); i++){
iss >> OGword;
totalWords++;
}
if (OGword[0]=='a' || OGword[0]=='A' || OGword[0]=='e' || OGword[0]=='E' || OGword[0]=='i' || OGword[0]=='I' || OGword[0]=='o' || OGword[0]=='O' || OGword[0]=='u' || OGword[0]=='U'){
cout << OGword << "way" << " ";
totalLetters += (OGword.size() + 3);
}
else {
PLword = OGword.substr(index);
length = PLword.length();
PLword.insert(length, "ay");
PLword.insert(length, 1, OGword[index]);
PLword.erase(0, 1);
if (isupper(OGword[0])){
transform(PLword.begin(), PLword.end(), PLword.begin(), ::tolower);
(toupper(PLword[1]));
char upper;
upper = toupper(PLword[0]);
PLword.erase(0, 1);
cout << upper;
}
cout << PLword << " ";
totalLetters += PLword.size();
}
} while (iss);
totalChars = totalLetters + 1;
cout << "\n\nTotal words: " << totalWords << endl;
cout << "Total Letters: " << totalLetters << endl;
cout << "Total Characters: "<< totalChars << endl;
}

Answer 1

Problem
The core loop of the program looks like this (in pseudocode):
istringstream iss; // Contains line of text.
do {
string OGword;
get_OGword_and_count_totalWords(iss, OGword);
print_pig_latin_of_word(OGword);
} while (iss);
The loop runs as long as iss has not experienced an error. And in particular, iss does not experience an error until an extraction operation fails. So things happen in the loop like this:
OGword contains the last legitimate word on the line.
Print the last word.
The while clause is tested. iss is still good at this point because no error has occurred, even if iss is at the end of string.
Attempt to extract a word into OGword. This fails, and leaves OGword empty ("").
Print the Pig Latin version of "", which is "ay".
The while clause is tested. iss is in an error state, and the loop ends.
Fix
One possible fix out of many is to test iss for an error immediately after extracting a word.
std::istringstream iss; // Contains line of text
std::string OGword;
while (iss >> OGword) {
increment_word_total();
print_pig_latin_of_word(OGword);
}
In this version, the operation iss >> OGword returns iss, which is converted to bool. If there was an error during the immediately preceeding extraction, the loop ends without printing anything.
Other Advice
I think the best way to improve readability is to break the code up into smaller functions. For instance, take the if / else block that formats and prints the Pig Latin, and actually put it in a function:
int print_pig_latin_of_word_and_return_total_letters(string_view word);
Then, the code in that function can be further subdivided:
bool starts_with_vowel(std::string_view word);
int print_vowel_word_and_count_letters(std::string_view word);
int print_consonant_word_and_count_letters(std::string_view word);
int print_pig_latin_of_word_and_count_letters(std::string_view word) {
if (starts_with_vowel(word)) {
return print_vowel_word_and_count_letters(word);
} else {
return print_consonant_word_and_count_letters(word);
}
}
Odds and Ends
I would drop using namespace std and write all of the std library names as std::string, etc. This makes it clear which things are from the standard library.
The program has interesting behavior on input files that contain more than one line. There is a for loop that loops over phrase.size() which is number of input lines. This causes words to be skipped and totalWords to be incorrect.
This statement doesn't do anything, because the result of toupper is ignored:
(toupper(PLword[1]));