There are several issues with the code, not withstanding that this could be written using C++, not C with a sprinkling of C++ I/O..
Issue 1:
Since you're using c-style strings, any copying of string data will require function calls such as strcpy(), strncpy(), etc. You failed in following this advice in this code:
for( int y = 0; y < new_numwords; y++)
{
for( int i = 0; i < NUM_WORDS; i++)
{
if (!strcmp(SentenceArry[i], EMPTY[0]) == 0)
{
New_SentenceArry[y] = SentenceArry[i]; // This is wrong
New_WordCount[y] = WordCount[i];
y++;
}
}
}
You should be using strcpy(), not =
to copy strings.
strcpy(New_SentenceArry[y], SentenceArry[i]);
Issue 2:
You should allocate WORD_LENGTH for both the original and new arrays. The length of the strings is independent of the number of strings.
char** New_SentenceArry = new char*[new_numwords]; //declare pointer for the sentence
for( int i = 0; i < new_numwords; i++)
{
New_SentenceArry[i] = new char[new_numwords];
}
This should be:
char** New_SentenceArry = new char*[new_numwords]; //declare pointer for the sentence
for( int i = 0; i < new_numwords; i++)
{
New_SentenceArry[i] = new char[WORD_LENGTH];
}
Issue 3:
Your loops do not check to see if the index is going out of bounds of your arrays.
It seems that you coded your program in accordance to the data that you're currently using, instead of writing code regardless of what the data will be. If you have limited yourself to 17 words, where is the check to see if the index goes above 16? Nowhere.
For example:
while (!read_text.eof() )
Should be:
while (!read_text.eof() && word_entry < NUM_WORDS)
Issue 4:
You don't process the first string found correctly:
read_text >> SentenceArry[word_entry]; // Here you read in the first word
while (!read_text.eof() )
{
word_entry++; //increment counter
read_text >> SentenceArry[word_entry]; // What about the first word you read in?
Summary:
Even with these changes, I can't guarantee that the program won't crash. Even it it doesn't crash with these changes, I can't guarantee it will work 100% of the time -- a guarantee would require further analysis.
The proper C++ solution, given what this assignment was about, is to use a std::map<std::string, int>
to keep the word frequency. The map would automatically store similar words in one entry (given that you remove the junk from the word), and would bump up the count to 1 automatically, when the entry is inserted into the map.
Something like this:
#include <string>
#include <map>
#include <algorithm>
typedef std::map<std::string, int> StringMap;
using namespace std;
bool isCharacterGarbage(char ch)
{ return ch == ',' || ch == '.'; }
int main()
{
StringMap sentenceMap;
//...
std::string temp;
read_text >> temp;
temp.erase(std::remove_if(temp.begin(), temp.end(), isCharacterGarbage),temp.end());
sentenceMap[temp]++;
//...
}
That code alone does everything your original code did -- keep track of the strings, bumps up the word count, removes the junk characters from the word before being processed, etc. But best of all, no manual memory management. No calls to new[], delete[], nothing. The code just "works". That is effectively 5 lines of code that you would just need to write a "read" loop around.
I won't go through every detail, you can do that for yourself since the code is small, and there are vast amounts of resources available explaining std::map
, remove_if()
, etc.
Then printing out is merely going through the map and printing each entry (string and count). If you add the printing, that may be 4 lines of extra code. So in all, practically all of the assignment is done with effectively 10 or so lines of code.