How to access values from pointers to pointers in C?

Question

I am writing a program that is attempting to implement a hash table in C. It has to use a pointer to a pointer as an attribute in a struct data type. Here is the code of the relevant structs.

      struct node {
      char * data;
      struct node * next;
    };

    struct hashtable {
      int size;
      struct node ** table;
    };

Part one of question:

I have been doing a lot of research to figure out just how pointers to pointers work but I still can't seem to get my head around it. Every example I've found assigns two different pointer names to the first and second pointer. eg

  x = 5;
  *p = &x;
  **q = &p; 

What about the case above "struct node ** table;" Is ** table the same as **q in this case? What would the values of q , *q and **q be in this case? is q = 5 and then *q and **q work back through the addresses or is **q = 5 and *q and q store the addresses?

Part two of question:

How do I access a pointer to a pointer within another function? Here is the code I have right now.

struct hashtable * hashtable_new(int size){

    struct hashtable *newTable;

    if ( size < 1 )
    {
        return NULL;
    printf("Its returning Null.");
    }
    else
    {
        newTable = malloc(sizeof(struct hashtable));
        newTable -> size = size;
        newTable -> table = malloc(sizeof(struct node) * size);
        int i;
        for (i = 0; i < size; i++ )
        {

            newTable -> table[i] = NULL;
        }
    fprintf(stderr, "Table has been created.");
        return newTable;
    }

};

I'm not sure I understand how to access either the pointer or the pointer to a pointer through the -> symbol. Is it "newtable -> table" or "newtable -> -> table" ?

The aim of the table is essentially to be a 2d table, where the list is primarily 1D and can spread to 2D to deal with collisions in the hashing.

End note:

Hopefully I've provided enough information to make contextual sense asking this question. Its my first time asking a question on stackoverflow so feel free to ask me extra question, provide CC or flag any mistakes I've made asking this question.

Thank you!

Answer 1

There are several ways to look at a pointer to a pointer. One way is as you described with

int   x =  5;
int  *p = &x;
int **q = &p; 

After all that, the following are all true:

**q == *p ==  x == 5
 *q ==  p == &x
  q == &p

However, another way to look at a pointer to a pointer is as an array of pointers. For example, assume the declaration:

int *arr[10];

Except when it is the operand of the sizeof or unary & operands, the expression arr will be converted ("decay") from type "10-element array of int * " to "pointer to int * ", or int ** .

If you want to dynamically allocate an N -element array of pointers to struct node , you would write

struct node **arr = malloc( sizeof *arr * N );

You would use the [] subscript operator to access elements in arr as you would a normal array. Since arr has type struct node ** , each arr[i] has type struct node * , so you would access members of the structure like so:

arr[i]->data = "foo"; // use -> when the LHS is a pointer to a struct or union
arr[i]->next = NULL;

struct hashtable allows you to create multiple hashtable instances, each one of which has a size member and an array of struct node * which you allocate at runtime, like so:

struct hashtable h1;

h1.size = 512;
h1.table = calloc( h1.size, sizeof h1.table ); // calloc initializes all elements of
                                               // h1.table to NULL
if ( !h1.table )
{
  // calloc call failed, handle as appropriate
}

You'd insert a new element into the table like so:

/**
 * Assumes we know that the string data points to is not already in
 * the table.  hashfunc is a pointer to the hashing function.
 */
void insert( struct hashtable *h, const char *data, size_t (*hashfunc)(const char *))
{
  /**
   * Create a new table entry.  Remember that table is an array of
   * *pointers* to struct node, and that each array element is initially
   * NULL.  We still need to allocate something for that table entry
   * to point *to*.
   */
  struct node *newNode = malloc( sizeof *newNode );
  if ( !newNode )
  {
    /**
     * Memory allocation for new node failed, handle as appropriate
     */
  }

  newNode->data = malloc( strlen( data ) + 1 );
  strcpy( newNode->data, data );

  /**
   * Get the table index by computing the hash and modding against the
   * table size.
   */
  size_t idx = hashfunc( data ) % h->size;

  /**
   * Insert newNode at the head of the list starting at
   * h->table[idx].  In reality you'd probably want to insert 
   * strings in order, but for a toy example this is good enough.
   */
  newNode->next = h->table[idx]; // newNode->next points to the previous head of list
  h->table[idx] = newNode;       // newNode becomes new head of list.
}

When you're done, your hashtable looks something like this:

    +---+
h1: |   | size
    +---+             +---+     +---+---+     +---+---+
    |   | table ----> |   | --> |   |   | --> |   |   |
    +---+             +---+     +---+---+     +---+---+
                      |   | 
                      +---+     +---+---+
                      |   | --> |   |   |
                      +---+     +---+---+     +---+---+     +---+---+
                      |   | --> |   |   | --> |   |   | --> |   |   |
                      +---+     +---+---+     +---+---+     +---+---+
                       ...

table points to an array of struct node * , each of which may point to an instance of struct node , each of which may point to another instance of struct node . In the picture above, two strings have hashed to table[0] , one to table[2] , and three to table[3] .

Answer 2

Here's a little program that might help you, it might be worth playing about with pointers for a bit and inspecting them, output addresses etc.

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

int
main(void)
    {
        char    *p1 = "String of text.";
        char    **p2;

        /**
         * We can print the avtual address of *p1 and **p2...
         */
        fprintf(stdout, "p1 = %p\n", p1);
        fprintf(stdout, "p2 = %p\n", p2);

        /**
         * Now if we point p2 at p1 we can then check what address
         * *p2 points at...
         */
        p2 = &p1;
        fprintf(stdout, "*p2 = %p\n", *p2);

        /**
         * So if you think about it, *p2 should be the same at p1
         * if we print it as a string:
         */
        fprintf(stdout, "p1 is %s\n*p2 is %s\n", p1, *p2);

        exit(EXIT_FAILURE);
    }

Hope this helps.

As for referencing the struct from another function, if you're passing a pointer to that structure you use a single -> to refer to any member of the struct...but you are allocating another struct within your struct - using a pointer.

Say I have

struct one {
    char *string;
}

struct two {
    struct *one a;
    struct one b;
    char *whatever;
}

If have a pointer to a struct two:

struct two *t;

And it had its member a allocated a struct one...then I'd reference the members of a using ->

t -> a -> string;

But b isn't a pointer so you would do:

t -> b.string;