Programming

What bounds? Remember it's a big 1D array. Nothing stops you from going beyond the 'size'.

Also remember, 3 is TWO away from two. 0 isn't, 1 isn't, 2 is outside, 3 is outside.

No, I did not get you.
As per the question is to use (*pa)[1] = &a[1][0]; (*ppa)[3][2] = &a[1] to reach "r", what is the correct answer although multiple addresses can print it out? Need more digestion. Can you suggest a good reference book about this? Thanks a lot!

As Corona688 already said multidimensional arrays are nothing but a logical aggregation of 1D arrays. The data items are laid out sequentially so once you get that down it should be easy to visualize how they are combined logically to form 2 or 3 or n dimensional arrays. And the book I really like is "Expert C Programming" which has an entire chapter devoted to explaining multidimensional arrays along with lots of examples and pictures.

Thanks Shamrock!

I was trying to catch the "dynamic" of the pointer in this example. The part that tricks me is(*pa)[1] = &a[1][0]; (*ppa)[3][2] = &a[1].
I want to confirm the structure of (*pa)[1] and (*ppa)[3][2], which confused me with the assginment. &a[1][0] is the address for letter 'g',Is my understanding right? but what is &a[1]? I feel getting close to the point, but just not quite sure without a guide/reference/expert.
I read from the book saying the last dimension must always be given. I am not sure here for a[i][j][k], &a[1][0], i=1, j=0, k not assigned; and &a[1], i = 1, j, k not assigned. Or, (*pa)[1] is for k = 1; i, j not assigned here. Got totally lost. I wish the author had given explanation for this question.

Thanks!

Try printing its contents and see.

Thanks Corona688!
I fully understand the full array is actually a big 1-dimension array from your previous reply.
I tried:

Code:

printf("The value in a[1] is: %p\n", a[1]);
    printf("The value in a[0] is: %p\n", a[0]);
    printf("The value in a[1][0] is: %p\n", a[1][0]);
    printf("The value in a[1][1] is: %p\n", a[1][1]);

Code:

Output:
The value in a[1] is: 0x7fff80c346a6
The value in a[0] is: 0x7fff80c346a0
The value in a[1][0] is: 0x7fff80c346a6
The value in a[1][1] is: 0x7fff80c346a8

The value in a[1][0][0] is: g
The value in a[1][0][1] is: h

first 4 printf() print addresses in fact, and I seem understand that:
Because a[1] is the first address of a two-way (3x2) array, so a[1] is the same as a[1][0], which in turn is the first address of two-elements array, so that (now value, not address) a[1][0][0] = g; a[1][0][1] = h.
My confusion is: How is the connection with the pointer (*pa)[2]? Here the [2] is the last dimension of the array a[4][3][2]. I treat each "2-element array" as a box---a single unit. so that (*pa)[2] points a[4][3], each unit is an address of an array. Right?
I thought pa[4][3] should point to an address of the array{'q', 'r'}, but it actually points to the element, char "r". I lost the connection here (*pa)[2] vs. pa[4][3].
Another confusion is the declaration of (*pa)[2] = &a[1][0] and (*ppa)[3][2] = &a[1]. The author is to emphasize the address of array and pointer. By array address only one anwer for char "r" which is a[2][2][1]. With pointer, you can have many ways(!?) What is the trick to connect two of them, say, if I declare (*pa)[2] = &a[2][0]?

Sorry I am so slow with this important point.

Going back to your first post you have...

Code:

char a[4][3][2] = {...};    /* a 3-dimensional array */
char (*pa)[2] = &a[1][0];   /* pa is a pointer to array of 2 chars */
char (*ppa)[3][2] = &a[1];  /* ppa is a pointer to a 2-dimensional array of chars */

So a[4][3][2] is made up of 4 items...each item is an array of 3 arrays...each of those 3 arrays is an array of 2 chars.
From this you can see that the element a[1] is an array of 3 arrays each made up of an array of 2 chars
Effectively saying that a[1] is a 2-d array of rank 3 x 2 and &a[1] is a pointer to that aka (*ppa)[3][2].
&a[1][0] is a pointer to array of 2 chars...so it is a pointer to the array that contains 'g' and 'h'.
As a[1] is an array of 3 arrays each made up of an array of 2 chars...&a[1] is a pointer to that aka (*ppa)[3][2]
Omg! which book have you been reading
Read "last dimension must always be given" to say that only the innermost dimension can be omitted while the rest need to be given so that the compiler can allocate storage for the array.
Moreover this is only true when initializing the array...if the array is defined without initialization all dimensions must be given...