Question

Newbie here,

I have a struct for a word, which contains a char array for the words themselves(the struct has other functions, which are unrelated to my question) and I’m trying to store it in a hashmap, which is an array of word struct pointers. In my program, every time I see a new word, I create a new word struct and malloc the char-array to create it. However, after a few run through of the loop, it changes the old word to a new word, even though it’s at different hashmap locations.

What I’m wondering is if it’s possible to have the loop in which I create the new word struct point to a new address?

struct words add;
int b;
for(b = 0; b < strlen(LowerCaseCopy); b++)
{
    add.word[b] = '\0';
}
for(b=0;b< strlen(LowerCaseCopy);b++)
{
add.word[b] = LowerCaseCopy[b];
}
hashmap[hashf] = &add;

This is the code in question.

An example of my problem:
the first runthrough of the loop, I set add.word to apple, which is stored at a specific hashmap slot.
the next runthrough of the loop, I set add.word to orange, which is stored at a different slot. The problem is that at the first slot, it no longer stores apple, it instead stores orange, so I have 2 slots that store orange, which is not what I want. How do I fix this?

Answer 1

A simple solution (I think) would be to put the functionality to add entries to the hashmap in a separate function. This function allocates a new words structure and puts that in the hashmap:

void add_to_hashmap(struct something *hashmap, char *lower_case_word)
{
    /* Using "calloc" we don't have to manually clear the structure */
    struct words *words = calloc(1, sizeof(struct words));

    /* Copy +1 to include the terminating '\0' */
    memcpy(words->word, lower_case_word, strlen(lower_case_word) + 1);

    /* Replace this with whatever you use to calculate the hash */
    int hashf = calculate_hash(lower_case_word);

    hashmap[hashf] = words;
}

If you remove an entry (i.e. setting it to NULL) you have to remember to free it first.