Anagram

#include <stdio.h>
#include <stdlib.h>

#define MAXN 100000 // assume that maximum number of words is MAXN
#define ALLCHAR 256 // assume that maximum distinct alphabets is about ALLCHAR
#define MAXCHAR 256 // assume that maximum length is MAXCHAR/3
#define TRUE 1
#define FALSE 0
#define FILENAME "thaiwords.txt"

typedef struct{
    int bucket[ALLCHAR], id;
}Group;

typedef char bool;

char word[MAXN][MAXCHAR];
Group group[MAXN];

int cmp(const void* aa, const void* bb){
    Group* a = (Group*)aa;
    Group* b = (Group*)bb;
    int i;
    for(i = 0; i < ALLCHAR; ++i)
        if(a->bucket[i] != b->bucket[i]) return a->bucket[i] - b->bucket[i];

    return 0;
}

bool is_equal(int a, int b){
    int i;
    for(i = 0; i < ALLCHAR; ++i)
        if(group[a].bucket[i] != group[b].bucket[i]) return FALSE;

    return TRUE;
}

int main(){
    int n = 0, i, len;
    FILE* fp = fopen(FILENAME, "r");
    if(!fp) printf("File not found!\n"), exit(1);
    while(fgets(word[n], MAXCHAR,fp)){
        for(i = 0; word[n][i] != 0; ++i){
            if(word[n][i] == ' ') continue;
            group[n].bucket[(unsigned char)word[n][i]]++;
        }
        group[n].id = n;
        n++;
    }
    qsort(group, n, sizeof(group[0]), cmp);
    int allgroup = 0;

    printf("%s", word[group[0].id]); // note that fgets() get '\n' character.
    allgroup = 1;
    for(i = 1; i < n; ++i){
        if(is_equal(i - 1, i)){
            printf("%s", word[group[i].id]);
        }else{
            printf("\n");
            printf("%s", word[group[i].id]);
            allgroup++;
        }
    }
    printf("\nall groups = %d, all words = %d\n", allgroup, n);
    fclose(fp);
    return 0;
}

/*
Complexity:
O((MAXN * ALLCHAR * log(MAXN)) + (MAXN * MAXCHAR))
It is good to use this algorithm when ALLCHAR is low, but MAXCHAR is high.

With suffix tree or string hashing, this bound would be
O((MAXN * log(MAXN) * log(MAXCHAR)) + (MAXN * MAXCHAR * log(MAXCHAR))).
It is good to use this algorithm when ALLCHAR is high, but MAXCHAR is low.

Note:
From http://www.acc.umu.se/~saasha/charsets/?charset=iso-8859-11&charset=iso-8859-11
Each Thai character becomes three numbers.
'ก' = 224, 184, 129
'ข' = 224, 184, 130
...
'๛' = 224, 185, 155

Luckily, although we separate a single character to three numbers, there will not be the vague case; that is, we don't have to concern about false positive outcome.

Testcase:
sadadsasdasddsaasdasdasdads
สวัสดีabc
asdadsasdasdasdasdasdasdads
acสวีbดัส
abc
สวัสดีดี
asdadsasdasddsaasdasdasdads
*/