I implemented this class to create a trie data structure. The function
unsigned long Insert(string) //inserts the string in trie & return no of words in trie
void PrintAllWords(); // prints all words in trie separated by space in dictionary order
implementation works correctly and prints all the words inserted from a text file of english dictionary words when the number of words is not very large, but when supplied with a file with some 350k words it only prints out a b c d upto z.
private variables
struct TrieTree
{
std::map<char,struct TrieTree*> map_child;
std::map<char,unsigned long> map_count; //keeps incrementing count of char in map during insertion.
bool _isLeaf=false; // this flag is set true at node where word ends
};
struct TrieTree* _root=NULL;
unsigned long _wordCount=0;
unsigned long _INITIALIZE=1;
Below is complete implementation with driver program. The program is executable.
#include<iostream>
#include<map>
#include<fstream>
class Trie
{
private:
struct TrieTree
{
std::map<char,struct TrieTree*> map_child;
std::map<char,unsigned long> map_count;
bool _isLeaf=false;
};
struct TrieTree* _root=NULL;
unsigned long _wordCount=0;
unsigned long _INITIALIZE=1;
struct TrieTree* getNode()
{
return new TrieTree;
};
void printWords(struct TrieTree* Tptr,std::string pre)
{
if(Tptr->_isLeaf==true)
{
std::cout<<pre<<" ";
return;
}
std::map<char,struct TrieTree*>::iterator it;
it=Tptr->map_child.begin();
while(it!=Tptr->map_child.end())
{
pre.push_back(it->first);
printWords(it->second,pre);
pre.erase(pre.length()-1); //erase last prefix character
it++;
}
}
public:
Trie()
{
_root=getNode();
}
unsigned long WordCount()
{
return _wordCount;
}
unsigned long WordCount(std::string pre) //count words with prefix pre
{
if(WordCount()!=0)
{
struct TrieTree *Tptr=_root;
std::map<char,unsigned long>::iterator it;
char lastChar;
for(int i=0;i<pre.length()-1;i++)
{
Tptr=Tptr->map_child[pre[i]];
}
lastChar=pre[pre.length()-1];
it=Tptr->map_count.find(lastChar);
if(it!=Tptr->map_count.end())
{
return Tptr->map_count[lastChar];
}
else
{
return 0;
}
}
return 0;
}
unsigned long Insert(std::string key) //return word count after insertion
{
struct TrieTree *Tptr =_root;
std::map<char,struct TrieTree*>::iterator it;
if(!SearchWord(key))
{
for(int level=0;level<key.length();level++)
{
it=Tptr->map_child.find(key[level]);
if(it==Tptr->map_child.end())
{
//alphabet does not exist in map
Tptr->map_child[key[level]]=getNode(); // new node with value pointing to it
Tptr->map_count[key[level]] = _INITIALIZE;
Tptr=Tptr->map_child[key[level]]; //assign pointer to newly obtained node
if(level==key.length()-1)
Tptr->_isLeaf=true;
}
else
{ //alphabet exists at this level
Tptr->map_count[key[level]]++;
Tptr=Tptr->map_child[key[level]];
}
}
_wordCount++;
}
return _wordCount;
}
bool SearchWord(std::string key)
{
struct TrieTree *Tptr =_root;
std::map<char,struct TrieTree*>::iterator it;
for(int level=0;level<key.length();level++)
{
it=Tptr->map_child.find(key[level]);
// cout<<" "<<Tptr->map_child.size()<<endl; //test to count entries at each map level
if(it!=Tptr->map_child.end())
{
Tptr=Tptr->map_child[key[level]];
}
else
{
return false;
}
}
if(Tptr->_isLeaf==true)
return true;
return false;
}
void PrintAllWords()
{ //print all words in trie in dictionary order
struct TrieTree *Tptr =_root;
if(Tptr->map_child.empty())
{
std::cout<<"Trie is Empty"<<std::endl;
return;
}
printWords(Tptr,"");
}
void PrintAllWords(std::string pre)
{ //print all words in trie with prefix pre in Dictionary order
struct TrieTree *Tptr =_root;
if(Tptr->map_child.empty())
{
std::cout<<"Trie is Empty"<<std::endl;
return;
}
for(int i=0;i<pre.length();i++)
{
Tptr=Tptr->map_child[pre[i]];
}
printWords(Tptr,pre);
}
};
int main(){
Trie t;
std::string str;
std::fstream fs;
fs.open("words.txt",std::ios::in);
while(fs>>str){
t.Insert(str);
}
t.PrintAllWords();
return 0;
}
I don't understand the output, please take a look at the code and suggest a fix. Thanks
When you add the word "a", if there is no word starting with 'a' in the tree, you will add a "leaf" node with 'a' as the value. If you then add a word starting with 'a', such as "an", you will add the 'n' node as a child of the 'a' node. However, when you print all the words, you stop recursing when you hit a leaf node, meaning you ignore all the other words starting with that word.
Simple solution: remove the return from printWords.
Similarly if you already have "an" in the tree, when you add 'a', you don't mark it as a leaf, so it will never be output.
Simple solution: Set _isLeaf when adding a word, even if the node already exists (i.e. add Tptr->_isLeaf=true; to the else clause in Insert
I think you would be better off changing _isLeaf to something like _isWord as it seems odd to have leaf nodes with child items.