Finding all adjacent elements in a 2D array - c

I am working on a project where at one point I am stuck.
My question is for example I have the following 2D array containing 3 different integers.
2 2 2 2 1
1 2 2 2 1
3 3 2 3 2
3 1 3 3 1
1 1 2 3 1
1 3 1 3 3
What I want is to find the longest adjacent elements chain of array of any number contained in the array.
Like in the above array the longest chain is of digit 2.
2 2 2 2
2 2 2
2
Can anyone just guide me as to what I have to do to achieve this goal?

Easier to draw than to explain...
2 2 2 2 1 => A A A A B => (A: 4, B: 1)
1 2 2 2 1 => C A A A B => (A: 3 + 4, B: 1 + 1, C: 1)
3 3 2 3 2 => D D A E F => (A: 1 + 7, B: 2, C: 1, D: 2, E: 1, F: 1)
3 1 3 3 1 => D G E E G => (A: 8, B: 2, C: 1, D: 2 + 1, E: 2 + 1, F: 1, G: 1)
1 1 2 3 1 => ...
1 3 1 3 3 => ...
update:
And now, with some real code:
#include <stdlib.h>
#include <string.h>
#include <stdio.h>
#define ROWS 6
#define COLS 5
unsigned char eles[ROWS][COLS] = { { 2, 2, 2, 2, 1 },
{ 1, 2, 2, 2, 1 },
{ 3, 3, 2, 3, 2 },
{ 3, 1, 3, 3, 1 },
{ 1, 1, 2, 3, 1 },
{ 1, 3, 1, 3, 3 } };
struct zone {
int acu;
int row, col;
int refs;
};
typedef struct zone zone;
zone *
new_zone(int row, int col) {
zone *z = (zone *)malloc(sizeof(zone));
z->col = col;
z->row = row;
z->refs = 1;
z->acu = 0;
}
void croak (const char *str) {
fprintf(stderr, "error: %s\n", str);
exit(1);
}
void
free_zone(zone *z) {
if (z->refs != 0) croak("free_zone: reference count is not cero");
free(z);
}
zone *
ref_zone(zone *z) {
z->refs++;
return z;
}
void
unref_zone(zone *z) {
z->refs--;
if (!z->refs) free_zone(z);
}
int
main() {
zone *last[COLS];
zone *current[COLS];
zone *best = new_zone(0, 0);
int i, j;
memset(last, 0, sizeof(last));
for (j = 0; j < ROWS; j++) {
for (i = 0; i < COLS; i++) {
unsigned int ele = eles[j][i];
zone *z;
/* printf("analyzing ele: %d at row %d, col: %d\n", ele, j, i); */
if (i && (ele == eles[j][i-1])) {
/* printf(" equal to left element\n"); */
z = ref_zone(current[i-1]);
if (j && (ele == eles[j-1][i])) {
zone *z1 = last[i];
/* printf(" equal to upper element\n"); */
if (z != z1) {
int k;
/* printf(" collapsing zone %p\n", z1); */
z->acu += z1->acu;
for (k = 0; k < COLS; k++) {
if (last[k] == z1) {
last[k] = ref_zone(z);
unref_zone(z1);
}
}
for (k = 0; k < i; k++) {
if (current[k] == z1) {
current[k] = ref_zone(z);
unref_zone(z1);
}
}
}
}
}
else if (j && (ele == eles[j-1][i])) {
/* printf(" equal to upper element\n"); */
z = ref_zone(last[i]);
}
else {
/* printf(" new element\n"); */
z = new_zone(j, i);
}
z->acu++;
current[i] = z;
/* printf(" element zone: %p\n", z); */
}
for (i = 0; i < COLS; i++) {
if (j) unref_zone(last[i]);
last[i] = current[i];
if (best->acu < current[i]->acu) {
unref_zone(best);
best = ref_zone(current[i]);
/* printf("best zone changed to %p at row; %d, col: %d, acu: %d\n", best, best->row, best->col, best->acu); */
}
}
}
printf("best zone is at row: %d, col: %d, ele: %d, size: %d\n", best->row, best->col, eles[best->row][best->col], best->acu);
}

Suppose your matrix is a graph, and the elements are vertices. Two vertices are connected if they are adjacent and have the same value. If you perform any search in that graph, be it Breadth-First Search or Depth-First Search, you'll get exactly what you want. HTH

You could treat this like a picture in a paint application. Perform a flood-fill on each element in your 2D array (unless its filled already by something else) and keep track how many pixels you filled in each step.
If your array is declared like
int elements[5][5];
Then introduce a second array which tells whether you filled an element already (if you like, use a different type like bool if thats's okay in your C program):
int pixelFilled[5][5];
memset( pixelFilled, 0, sizeof( pixelFilled ) );
Next, write a recursive function which performs a flood fill and returns the numbers of elements which were filled (I'm writing this from the top of my head, no guarantee whatsoever that this function works as it is):
int floodFill( int x, int y ) {
int filledPixels = 0;
if ( !pixelFilled[x][y] ) {
++filledPixels;
pixelFilled[x][y] = 1;
} else {
return 0;
}
if ( x < 4 && elements[x+1][y] == elements[x][y] )
filledPixels += floodFill( x + 1, y );
if ( x > 0 && elements[x-1][y] == elements[x][y] )
filledPixels += floodFill( x - 1, y );
if ( y < 4 && elements[x][y+1] == elements[x][y] )
filledPixels += floodFill( x, y + 1 );
if ( y > 0 && elements[x][y-1] == elements[x][y] )
filledPixels += floodFill( x, y - 1 );
return filledPixels;
}
Finally, iterate over your array and try to fill it completely. Keep track of the largest filled array:
int thisArea = 0;
int largestArea = 0;
int x, y;
for ( y = 0; y < 5; ++y ) {
for ( x = 0; x < 5; ++x ) {
thisArea = floodFill( x, y );
if (thisArea > largestArea ) {
largestArea = thisArea;
}
}
}
Now, largestArea should contain the size of the longest chain of adjacent elements.

define another 2d array of the same size, initialize all cells to 0
set maxval to 0
if helper array is full of 1's go to 5, otherwise find a cell with 0 and do:
3.1 change value of the cell to 1
3.2 set a counter to 1
3.3 check all adjacent cells, if they're 0 in the helper array and the same value as current cell in the input array then counter++ and go to 2.1 with new coordinates.
maxval = max(maxval,counter), go to 3
return maxval
steps 3.1-3.3 should be implemented as a recursive function which takes coordinate and both arrays as arguments and returns 1+the sum of the returned values from the recursive calls.

I love this kind of problems :-) so here it is my answer.
As said by Frerich Raabe, this can be solved with a floodFill function. For example, opencv library would provide such a function off the shelf.
Please forgive me if in the following code you'll find traces of C++, in case they should be simple to be replaced.
typedef struct Point {
int x;
int y;
} Point;
int areaOfBiggestContiguousRegion(int* mat,int nRows, int nCols) {
int maxArea = 0;
int currValue, queueSize, queueIndex;
int* aux;
Point queue[1000]; //Stores the points I need to label
Point newPoint, currentPoint;
int x,y,x2,y2;
//Code: allocate support array aux of same size of mat
//Code: fill aux of zeros
for (y = 0; y < nRows; y++)
for (x = 0; x < nCols; x++)
if (aux[y * nCols + x] == 0) {//I find a pixel not yet labeled, my seed for the next flood fill
queueIndex = 0; //Contains the index to the next element in the queue
queueSize = 0;
currValue = mat[y * nCols + x]; //The "color" of the current spot
aux[y * nCols + x] = 1;
newPoint.x = x;
newPoint.y = y;
queue[queueSize] = newPoint;
queueSize++;
while(queueIndex != queueSize) {
currPoint = queue[queueIndex];
queueIndex++;
//Look left, right, up, down
x2 = currPoint.x - 1;
y2 = currPoint.y;
//Some copy & paste, sorry I have been too long on C++ to remember correctly about C functions
if (x2 >= 0 && aux[y2 * nCols + x2] == 0 && mat[y2 * nCols + x2] == currValue) {
aux[y2 * nCols + x2] = 1;
newPoint.x = x2;
newPoint.y = y2;
queue[queueSize] = newPoint;
queueSize++;
}
x2 = currPoint.x + 1;
y2 = currPoint.y;
//Some copy & paste, sorry I have been too long on C++ to remember correctly about C functions
if (x2 < nCols && aux[y2 * nCols + x2] == 0 && mat[y2 * nCols + x2] == currValue) {
aux[y2 * nCols + x2] = 1;
newPoint.x = x2;
newPoint.y = y2;
queue[queueSize] = newPoint;
queueSize++;
}
x2 = currPoint.x;
y2 = currPoint.y - 1;
//Some copy & paste, sorry I have been too long on C++ to remember correctly about C functions
if (y2 >= 0 && aux[y2 * nCols + x2] == 0 && mat[y2 * nCols + x2] == currValue) {
aux[y2 * nCols + x2] = 1;
newPoint.x = x2;
newPoint.y = y2;
queue[queueSize] = newPoint;
queueSize++;
}
x2 = currPoint.x;
y2 = currPoint.y + 1;
//Some copy & paste, sorry I have been too long on C++ to remember correctly about C functions
if (y2 < nRows && aux[y2 * nCols + x2] == 0 && mat[y2 * nCols + x2] == currValue) {
aux[y2 * nCols + x2] = 1;
newPoint.x = x2;
newPoint.y = y2;
queue[queueSize] = newPoint;
queueSize++;
}
} //while
if (queueSize > maxArea)
maxArea = queueSize; //If necessary we could store other details like currentValue
}//if (aux...
return maxArea;
}
Note: In C++ using std containers and a constructor for Point it becomes much more compact

Related

Fastest way to compute maximal n s.t. n over k <= x

I'm looking for a fast way to compute the maximal n s.t. n over k <= x for given k and x.
In my context n \leq n' for some known constant n', lets say 1000. k is either 1,2, or 3 and x is choosen at random from 0 ... n' over k
My current approach is to compute the binomial coefficient iterativly, starting from a_0 = k over k = 1. The next coefficient a_1 = k+1 over k can be computed as a_1 = a_0 * (k+1) / 1 and so on.
The current C code looks like this
uint32_t max_bc(const uint32_t a, const uint32_t n, const uint32_t k) {
uint32_t tmp = 1;
int ctr = 0;
uint32_t c = k, d = 1;
while(tmp <= a && ctr < n) {
c += 1;
tmp = tmp*c/d;
ctr += 1;
d += 1;
}
return ctr + k - 1;
}
int main() {
const uint32_t n = 10, w = 2;
for (uint32_t a = 0; a < 10 /*bc(n, w)*/; a++) {
const uint32_t b = max_bc(a, n, w);
printf("%d %d\n", a, b);
}
}
which outputs
0 1
1 2
2 2
3 3
4 3
5 3
6 4
7 4
8 4
9 4
So I'm looking for a Bittrick or something to get around the while-loop to speed up my application. Thats because the while loop gets executedat worst n-k times. Precomputation is not an option, because this code is part of a bigger algorithm which uses a lot of memory.
Thanks to #Aleksei
This is my solution:
template<typename T, const uint32_t k>
inline T opt_max_bc(const T a, const uint32_t n) {
if constexpr(k == 1) {
return n - k - a;
}
if constexpr (k == 2) {
const uint32_t t = __builtin_floor((double)(__builtin_sqrt(8 * a + 1) + 1)/2.);
return n - t - 1;
}
if constexpr (k == 3) {
if (a == 1)
return n-k-1;
float x = a;
float t1 = sqrtf(729.f * x * x);
float t2 = cbrtf(3.f * t1 + 81.f * x);
float t3 = t2 / 2.09f;
float ctr2 = t3;
int ctr = int(ctr2);
return n - ctr - k;
}
if constexpr (k == 4) {
const float x = a;
const float t1 = __builtin_floorf(__builtin_sqrtf(24.f * x + 1.f));
const float t2 = __builtin_floorf(__builtin_sqrtf(4.f * t1 + 5.f));
uint32_t ctr = (t2 + 3.f)/ 2.f - 3;
return n - ctr - k;
}
// will never happen
return -1;
}
If k is really limited to just 1, 2 or 3, you can use different methods depending on k:
k == 1: C(n, 1) = n <= x, so the answer is n.
k == 2: C(n, 2) = n * (n - 1) / 4 <= x. You can solve the equation n * (n - 1) / 4 = x, the positive solution is n = 1/2 (sqrt(16x + 1) + 1), the answer to the initial question should be floor( 1/2 (sqrt(16x + 1) + 1) ).
k == 3: C(n, 3) = n(n-1)(n-2)/6 <= x. There is no nice solution, but the formula for the number of combinations is straightforward, so you can use a binary search to find the answer.

how to print a matrix with a negated row more efficiently

I wrote the following function:
void negate_row(const int n, const int r, int *a)
{
if (r == 0)
{
printf("Matrix with negated row: ");
printf("\n");
for (int y = 0; y < 3; y++)
{
*(a + 3 * r + y) = *(a + 3 * r + y) *(-1);
printf("%d ", *(a + 3 * r + y));
}
printf("\n");
for (int y = 0; y < 3; y++)
{
*(a + 3 * r + y) = *(a + 3 * r + y) *(-1);
printf("%d ", *(a + 3 * (r + 1) + y));
}
printf("\n");
for (int y = 0; y < 3; y++)
{
*(a + 3 * r + y) = *(a + 3 * r + y) *(-1);
printf("%d ", *(a + 3 * (r + 2) + y));
}
printf("\n");
}
So basically what happens here is my function takes in a n X 3 matrix and negates a specific row using pointer arithmetic. I have been able to achieve that, I've also been able to figure out how to print that same matrix with the negated row. It's just the way I'm doing it is not efficient at all. I'd have to write an if statement for each row, ex if r == 0,1,2,3,4 etc... is there any way I can do this more efficiently?
Some clarifications: const int n decides the size of the matrix (n x 3), const int r decides what row is negated (0 <= r < n).
A second loop will help. I generally find pointer code a bit harder to read. Particularly with Matrix manipulation you might be better off using array syntax instead of pointer syntax.
for (int y = 0; y < 3; y++)
{
for (int x = 0; x < 3; x++)
{
printf("%d ", *(a + 3 * (r + x) + y));
}
printf("\n");
}
Particular case (3 rows):
int mul = 1, y = 0;
for (int x = 0; x < n; x++) {
mul = x == r ? -1 : 1;
y = (a + 3 * x);
printf("%d ", (*y) * mul);
printf("%d ", (*y + 1) * mul);
printf("%d ", (*y + 2) * mul);
printf("\n");
}
More generic (m rows):
int mul = 1;
for (int x = 0; x < n; x++) {
mul = x == r ? -1 : 1;
for (int y = 0; y < m; y++) {
printf("%d ", ((*(a + m * x + y)) * mul);
}
printf("\n");
}
Note: In new compilers there is also no difference in speed between array or pointer syntax.
You could write a function that accepts a one-dimensional array (a row of your matrix) and then using this function you could in a loop output its all rows or output a selected row.
Here is a demonstrative program.
#include <stdio.h>
void negate_row( const int *a, size_t n, int width )
{
if ( width < 1 ) width = 1;
for ( const int *p = a; p != a + n; ++p )
{
printf( "%*d ", width, -*p );
}
putchar( '\n' );
}
int main(void)
{
enum { M = 3, N = 4 };
int matrix[M][N] =
{
{ 0, 1, 2, 4 },
{ 5, 6, 7, 8 },
{ 9, 10, 11, 12 }
};
for ( int ( *p )[N] = matrix; p != matrix + M; ++p )
{
negate_row( *p, N, 3 );
}
return 0;
}
The program output is
0 -1 -2 -4
-5 -6 -7 -8
-9 -10 -11 -12
As you showed a code where you are using pointers tp output elements of an array then in this demonstrative program I am also using pointers everywhere to access elements of an array.
The third parameter of the function specifies the width of the field for an outputted value.
To output the matrix in the reversed order of rows you can use the loop shown in the program blow.
#include <stdio.h>
void negate_row( const int *a, size_t n, int width )
{
if ( width < 1 ) width = 1;
for ( const int *p = a; p != a + n; ++p )
{
printf( "%*d ", width, -*p );
}
putchar( '\n' );
}
int main(void)
{
enum { M = 3, N = 4 };
int matrix[M][N] =
{
{ 0, 1, 2, 4 },
{ 5, 6, 7, 8 },
{ 9, 10, 11, 12 }
};
for ( int ( *p )[N] = matrix + M; p != matrix; )
{
negate_row( *--p, N, 3 );
}
return 0;
}
The program output is
-9 -10 -11 -12
-5 -6 -7 -8
0 -1 -2 -4
You can generalize the function easily: as long as the row is within the matrix, your code works for any row. Also note that it is better to use a separate function to print the matrix.
#include <stdio.h>
void negate_row(const int n, const int r, int *a) {
if (r >= 0 && r < n) {
// negate row r
for (int col = 0; col < 3; col++) {
*(a + 3 * r + col) *= -1;
}
}
}
void print_matrix(const int n, int *a, const char *title) {
if (title) {
printf("%s:\n", title);
}
for (int row = 0; row < n; row++) {
for (int col = 0; col < 3; col++) {
printf("%d ", *(a + 3 * row + col));
}
printf("\n");
}
printf("\n");
}
int main() {
int matrix[5 * 3] = {
0, 1, 2,
3, 4, 5,
6, 7, 8,
9, 10, 11,
12, 13, 14,
};
print_matrix(5, matrix, "Matrix");
negate_row(5, 0, matrix);
print_matrix(5, matrix, "Matrix with negated row");
negate_row(5, 3, matrix);
print_matrix(5, matrix, "Matrix with two negated rows");
negate_row(5, 0, matrix);
negate_row(5, 3, matrix);
print_matrix(5, matrix, "Matrix back to origin");
return 0;
}
Output:
Matrix:
0 1 2
3 4 5
6 7 8
9 10 11
12 13 14
Matrix with negated row:
0 -1 -2
3 4 5
6 7 8
9 10 11
12 13 14
Matrix with two negated rows:
0 -1 -2
3 4 5
6 7 8
-9 -10 -11
12 13 14
Matrix back to origin:
0 1 2
3 4 5
6 7 8
9 10 11
12 13 14

Code returning different outputs using goto

I'm writing a function that takes an input n and creates a 1D array of size (2n-1)^2 to simulate a square. I.e. for an input of n = 1, there is just one point, for an input of n = 2, it would look like
0 1 2
3 4 5
6 7 8
and for n = 3 it would look like
0 1 2 3 4
5 6 7 8 9
10 11 12 13 14
15 16 17 18 19
20 21 22 23 24
where each number is a point.
The function terminates when the current location is detected at being on an edge and the point attempts to move off of the square's grid.
The purpose of this is to simulate how many points are visited for squares of different sizes, from n=2^0 up to n=2^8 and to return the fraction of how many points are visited over the total amount of points in the square.
The function generates a random number and checks the modulus of it against 4, and if it returns 0, the location is moved up 1, if it returns 1, the location is moved right, 2 goes down, and 3 goes left.
#include <stdio.h>
#include <stdlib.h>
#include <math.h>
double two_d_random (int n) {
int tot_points = (2 * n - 1)*(2 * n - 1);
int value = (n*n)+((n-1)*(n-1))-1; //center
int length = 2 * n - 1; //length of side
int *array = (int *)malloc (sizeof (int) * tot_points);
int count = 0;
array[value] = 1;
while (1 == 1) {
int r = rand () % 4;
array[value] = 1;
if (r == 0) {//UP
if ((value >= 0) && (value < length)) {
goto a;
}
else {
array[value] = 1;
value -= length;
}
}
else if (r == 1) {//RIGHT
if ((value % length) == (2*n-2)){
goto a;
}
else {
array[value] = 1;
value += 1;
}
}
else if (r == 2) {//DOWN
if ((value < tot_points) && (value >= (tot_points - length))) {
goto a;
}
else {
array[value] = 1;
value += length;
}
}
else if (r == 3) {//LEFT
if (value % length == 0) {
goto a;
}
else {
array[value] = 1;
value -= 1;
}
}
}
a:
for (int i = 0; i < tot_points; i++) {
if (array[i] == 1) {
count += 1;
}
}
free (array);
return 1.0 * count / tot_points;
}
int main ()
{
int trials = 1000;
srand (12345);
for (int n = 1; n <= 256; n *= 2)
{
double sum = 0.;
for (int i = 0; i < trials; i++)
{
double p = two_d_random(n);
sum += p;
}
printf ("%d %.3lf\n", n, sum / trials);
}
return 0;
}
My current problem is that while I run it on my machine, I get a range of values I'm not expecting:
However, when a colleague runs it on their machine, they get the following which is what I expected:
I realize this is a large amount to ask at one point. I also realize that I should not use goto. I've put a lot of time into this however and I've no idea how to fix this. Any help is greatly appreciated.
You need to initialise your array. When calling malloc() it just returns a chunk of un-initialised memory. Either initialise it, or use calloc() instead, to get pre-zeroed memory.
double two_d_random( int n )
{
int tot_points = ( 2 * n - 1 ) * ( 2 * n - 1 );
int value = ( n * n ) + ( ( n - 1 ) * ( n - 1 ) ) - 1; //center
int length = 2 * n - 1; //length of side
int *array = (int *) malloc( sizeof( int ) * tot_points );
int count = 0;
// Initialise the array to zero
for ( int i=0; i<tot_points; i++ )
{
array[i] = 0;
}
array[value] = 1;
while ( 1 == 1 )
{
int r = rand() % 4;
array[value] = 1;
if ( r == 0 )
With this modification, I get results similar to what you reported as desired:
1 1.000
2 0.367
4 0.221
8 0.154
16 0.122
32 0.101
64 0.085
128 0.077
256 0.071

Swap columns using pointers in C

I am really stucked on this problem, my C code has worked very well using multidimensional arrays but i need to do the same using pointers but i'll describe the problem first.
Having the following matrix, i will get a number which will be the number of permutations (the number of swapping of columns that will move to the right and the last column will move to the first column).
For example
The number of column's permutations: 5
| 1 2 3 | -----> | 2 3 1 |
| 3 1 2 | -----> | 1 2 3 |
| 2 3 1 | -----> | 3 1 2 |
I wrote the following code using pointers, as you can see i build the matrix with multidimensional array and assign all of it into a pointer:
short elementMatrix[3][3] = {{1, 2, 3},
{3, 1, 2},
{2, 3, 1}};
short *element_matrix;
element_matrix = *elementMatrix;
int counter = 1;
while (counter <= 5)
{
for (int i = 0; i < 3; i++)
{
int temp = elementMatrix[i][PR.elem_mat_size - 1];
*outElementMatrix = *outElementMatrix + i * PR.elem_mat_size + PR.elem_mat_size - 1;
for (int j = 3 - 1; j >= 0; j--)
{
*(outElementMatrix + i * PR.elem_mat_size + j) = *(outElementMatrix + i * PR.elem_mat_size + j - 1);
if (j == 0)
{
*(outElementMatrix + i * PR.elem_mat_size + j) = *outElementMatrix;
}
}
}
counter++;
}
Since you want to swap out columns, it makes sense to have the pointers represent the columns. That way, you can swap a pointer to swap a column. So let's have an array of 3 pointers to a column.
short* col[3];
Each column consists of 3 shorts, so allocate that much memory.
for (int i = 0; i < 3; i++) {
col[i] = (short*)malloc(3 * sizeof(short));
}
Now to initialize the Matrix. This is a bit verbose, so if anyone knows a better way, edit away. :)
col[0][0] = 1; col[1][0] = 2; col[2][0] = 3;
col[0][1] = 3; col[1][1] = 1; col[2][1] = 2;
col[0][2] = 2; col[1][2] = 3; col[2][2] = 1;
Now we do the swap. Note how you need a temp variable, like Rishikesh Raje suggested. Also note that three swaps bring it back to the original, so instead of swapping n times, you only have to swap n % 3 times. Of course it's going to be pretty much instant with 5 or 2 swaps, but if you have to do like a billion, the difference should be noticeable.
for (int i = 0; i < 5; i++) {
short* temp = col[2];
col[2] = col[1];
col[1] = col[0];
col[0] = temp;
}
We assure that the result is correct by printing it:
for (int i = 0; i < 3; i++) {
printf("%d %d %d\n", col[0][i], col[1][i], col[2][i]);
}
You can consider the permutations as a rotation of each row in the matrix and, unless you have to somehow use the matrix after each step, calculate only the final result.
I'll use an extra buffer to help with the swaps.
#include <stdio.h>
#include <stddef.h>
#include <string.h>
#include <assert.h>
// rotate the values of an array using a buffer to ease the swappings
void rotate_(size_t n, void *arr, void *tmp, size_t offset)
{
assert(n && arr && tmp && offset <= n);
// casting to perform pointer arithmetic
memcpy(tmp, (char *)arr + (n - offset), offset);
memmove((char *)arr + offset, arr, n - offset);
memcpy(arr, tmp, offset);
}
void rotate_columns_short(size_t r, size_t c, short mat[r][c], short *buf, int n)
{
// clamp the value of the offset to the number of columns
size_t offset = (n >= 0
? n % c
: c - -n % c) * sizeof(short);
// transform each row
for (short *row = &mat[0][0], *row_end = row + r * c;
row != row_end;
row += c)
{
rotate_(c * sizeof(short), row, buf, offset);
}
}
void print_matrix_short(size_t r, size_t c, short mat[r][c])
{
for (size_t i = 0; i < r; ++i)
{
for (size_t j = 0; j < c; ++j)
{
printf(" %hd", mat[i][j]);
}
puts("");
}
}
#define ROWS 3
#define COLS 3
int main(void)
{
short matrix[ROWS][COLS] = {{1, 2, 3},
{3, 1, 2},
{2, 3, 1}};
short buf[COLS];
print_matrix_short(ROWS, COLS, matrix);
puts("");
rotate_columns_short(ROWS, COLS, matrix, buf, 5);
print_matrix_short(ROWS, COLS, matrix);
}
The output beeing:
1 2 3
3 1 2
2 3 1
2 3 1
1 2 3
3 1 2

Print two-dimensional array in spiral order

How do I print a 5×5 two-dimensional array in spiral order?
Is there any formula so that I can print an array of any size in spiral order?
The idea is to treat the matrix as a series of layers, top-right layers and bottom-left layers. To print the matrix spirally we can peel layers from these matrix, print the peeled part and recursively call the print on the left over part. The recursion terminates when we don't have any more layers to print.
Input matrix:
1 2 3 4
5 6 7 8
9 0 1 2
3 4 5 6
7 8 9 1
After peeling top-right layer:
1 2 3 4
8
5 6 7 2
9 0 1 6
3 4 5 1
7 8 9
After peeling bottom-left layer from sub-matrix:
6 7
5 0 1
9 4 5
3
7 8 9
After peeling top-right layer from sub-matrix:
6 7
1
0 5
4
After peeling bottom-left layer from sub-matrix:
0
4
Recursion terminates.
C functions:
// function to print the top-right peel of the matrix and
// recursively call the print bottom-left on the submatrix.
void printTopRight(int a[][COL], int x1, int y1, int x2, int y2) {
int i = 0, j = 0;
// print values in the row.
for(i = x1; i<=x2; i++) {
printf("%d ", a[y1][i]);
}
// print values in the column.
for(j = y1 + 1; j <= y2; j++) {
printf("%d ", a[j][x2]);
}
// see if more layers need to be printed.
if(x2-x1 > 0) {
// if yes recursively call the function to
// print the bottom left of the sub matrix.
printBottomLeft(a, x1, y1 + 1, x2-1, y2);
}
}
// function to print the bottom-left peel of the matrix and
// recursively call the print top-right on the submatrix.
void printBottomLeft(int a[][COL], int x1, int y1, int x2, int y2) {
int i = 0, j = 0;
// print the values in the row in reverse order.
for(i = x2; i>=x1; i--) {
printf("%d ", a[y2][i]);
}
// print the values in the col in reverse order.
for(j = y2 - 1; j >= y1; j--) {
printf("%d ", a[j][x1]);
}
// see if more layers need to be printed.
if(x2-x1 > 0) {
// if yes recursively call the function to
// print the top right of the sub matrix.
printTopRight(a, x1+1, y1, x2, y2-1);
}
}
void printSpiral(int arr[][COL]) {
printTopRight(arr,0,0,COL-1,ROW-1);
printf("\n");
}
Pop top row
Transpose and flip upside-down (same as rotate 90 degrees counter-clockwise)
Go to 1
Python 2 code:
import itertools
arr = [[1,2,3,4],
[12,13,14,5],
[11,16,15,6],
[10,9,8,7]]
def transpose_and_yield_top(arr):
while arr:
yield arr[0]
arr = list(reversed(zip(*arr[1:])))
print list(itertools.chain(*transpose_and_yield_top(arr)))
For python 3x:
import itertools
arr = [[1,2,3,4],
[12,13,14,5],
[11,16,15,6],
[10,9,8,7]]
def transpose_and_yield_top(arr):
while arr:
yield arr[0]
arr = list(reversed(list(zip(*arr[1:]))))
print(list(itertools.chain(*transpose_and_yield_top(arr))))
I see that no one has use only one for loop and without recursion in the code, and so I want to contribute.
The idea is like this:
Imagine there is a turtle standing at point (0,0), that is, top-left corner, facing east (to the right)
It will keep going forward and each time it sees a sign, the turtle will turn right
So if we put the turtle at point (0,0) facing right-ward, and if we place the signs at appropriate places, the turtle will traverse the array in spiral way.
Now the problem is: "Where to put the signs?"
Let's see where we should put the signs (marked by #, and numbers by O):
For a grid that looks like this:
O O O O
O O O O
O O O O
O O O O
We put the signs like this:
O O O #
# O # O
O # # O
# O O #
For a grid that looks like this:
O O O
O O O
O O O
O O O
We put the signs like this:
O O #
# # O
O # O
# O #
And for a grid that looks like this:
O O O O O O O
O O O O O O O
O O O O O O O
O O O O O O O
O O O O O O O
We put the signs like this:
O O O O O O #
# O O O O # O
O # O O # O O
O # O O O # O
# O O O O O #
We can see that, unless the point is at the top-left part, the signs are places at points where the distances to the closest horizontal border and the closest vertical border are the same, while for the top-left part, the distance to the top border is one more than the distance to the left border, with priority given to top-right in case the point is horizontally centered, and to top-left in case the point is vertically centered.
This can be realized in a simple function quite easily, by taking the minimum of (curRow and height-1-curRow), then the minimum of (curCol and width-1-curCol) and compare if they are the same. But we need to account for the upper-left case, that is, when the minimum is curRow and curCol themselves. In that case we reduce the vertical distance accordingly.
Here is the C code:
#include <stdio.h>
int shouldTurn(int row, int col, int height, int width){
int same = 1;
if(row > height-1-row) row = height-1-row, same = 0; // Give precedence to top-left over bottom-left
if(col >= width-1-col) col = width-1-col, same = 0; // Give precedence to top-right over top-left
row -= same; // When the row and col doesn't change, this will reduce row by 1
if(row==col) return 1;
return 0;
}
int directions[4][2] = {{0,1},{1,0},{0,-1},{-1,0}};
void printSpiral(int arr[4][4], int height, int width){
int directionIdx=0, i=0;
int curRow=0, curCol=0;
for(i=0; i<height*width; i++){
printf("%d ",arr[curRow][curCol]);
if(shouldTurn(curRow, curCol, height, width)){
directionIdx = (directionIdx+1)%4;
}
curRow += directions[directionIdx][0];
curCol += directions[directionIdx][1];
}
printf("\n");
}
int main(){
int arr[4][4]= {{1,2,3,4},{5,6,7,8},{9,10,11,12},{13,14,15,16}};
printSpiral(arr, 4, 4);
printSpiral(arr, 3, 4);
}
Which outputs:
1 2 3 4 8 12 16 15 14 13 9 5 6 7 11 10
1 2 3 4 8 12 11 10 9 5 6 7
Here are the three interesting ways
Reading in spiral way can be treated like a snake moving towards boundary and turning on hitting the boundary or itself (I find it elegant and most efficient being a single loop of N iterations)
ar = [
[ 0, 1, 2, 3, 4],
[15, 16, 17, 18, 5],
[14, 23, 24, 19, 6],
[13, 22, 21, 20, 7],
[12, 11, 10, 9, 8]]
def print_spiral(ar):
"""
assuming a rect array
"""
rows, cols = len(ar), len(ar[0])
r, c = 0, -1 # start here
nextturn = stepsx = cols # move so many steps
stepsy = rows-1
inc_c, inc_r = 1, 0 # at each step move this much
turns = 0 # how many times our snake had turned
for i in range(rows*cols):
c += inc_c
r += inc_r
print ar[r][c],
if i == nextturn-1:
turns += 1
# at each turn reduce how many steps we go next
if turns%2==0:
nextturn += stepsx
stepsy -= 1
else:
nextturn += stepsy
stepsx -= 1
# change directions
inc_c, inc_r = -inc_r, inc_c
print_spiral(ar)
output:
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
A recursive approach would be to print outer layer and call same function for inner rectangle e.g.
def print_spiral(ar, sr=0, sc=0, er=None, ec=None):
er = er or len(ar)-1
ec = ec or len(ar[0])-1
if sr > er or sc > ec:
print
return
# print the outer layer
top, bottom, left, right = [], [], [], []
for c in range(sc,ec+1):
top.append(ar[sr][c])
if sr != er:
bottom.append(ar[er][ec-(c-sc)])
for r in range(sr+1,er):
right.append(ar[r][ec])
if ec != sc:
left.append(ar[er-(r-sr)][sc])
print " ".join([str(a) for a in top + right + bottom + left]),
# peel next layer of onion
print_spiral(ar, sr+1, sc+1, er-1, ec-1)
Finally here is a small snippet to do it, not efficient but fun :), basically it prints top row, and rotates whole rectangle anti-clockwise and repeats
def print_spiral(ar):
if not ar: return
print " ".join(str(a) for a in ar[0]),
ar = zip(*[ reversed(row) for row in ar[1:]])
print_spiral(ar)
This program works for any n*n matrix..
public class circ {
public void get_circ_arr (int n,int [][] a)
{
int z=n;
{
for (int i=0;i<n;i++)
{
for (int l=z-1-i;l>=i;l--)
{
int k=i;
System.out.printf("%d",a[k][l]);
}
for (int j=i+1;j<=z-1-i;j++)
{
int k=i;
{
System.out.printf("%d",a[j][k]);
}
}
for (int j=i+1;j<=z-i-1;j++)
{
int k=z-1-i;
{
System.out.printf("%d",a[k][j]);
}
}
for (int j=z-2-i;j>=i+1;j--)
{
int k=z-i-1;
{
System.out.printf("%d",a[j][k]);
}
}
}
}
}
}
Hope it helps
I was obsessed with this problem when I was learning Ruby. This was the best I could do:
def spiral(matrix)
matrix.empty? ? [] : matrix.shift + spiral(matrix.transpose.reverse)
end
You can check out some of my other solutions by stepping back through the revisions in this gist. Also, if you follow the link back to whom I forked the gist from, you'll find some other clever solutions. Really interesting problem that can be solved in multiple elegant ways — especially in Ruby.
JavaScript solution:
var printSpiral = function (matrix) {
var i;
var top = 0;
var left = 0;
var bottom = matrix.length;
var right = matrix[0].length;
while (top < bottom && left < right) {
//print top
for (i = left; i < right; i += 1) {
console.log(matrix[top][i]);
}
top++;
//print right column
for (i = top; i < bottom; i += 1) {
console.log(matrix[i][right - 1]);
}
right--;
if (top < bottom) {
//print bottom
for (i = right - 1; i >= left; i -= 1) {
console.log(matrix[bottom - 1][i]);
}
bottom--;
}
if (left < right) {
//print left column
for (i = bottom - 1; i >= top; i -= 1) {
console.log(matrix[i][left]);
}
left++;
}
}
};
One solution involves directions right, left, up, down, and their corresponding limits (indices). Once the first row is printed, and direction changes (from right) to down, the row is discarded by incrementing the upper limit. Once the last column is printed, and direction changes to left, the column is discarded by decrementing the right hand limit... Details can be seen in the self-explanatory C code.
#include <stdio.h>
#define N_ROWS 5
#define N_COLS 3
void print_spiral(int a[N_ROWS][N_COLS])
{
enum {up, down, left, right} direction = right;
int up_limit = 0,
down_limit = N_ROWS - 1,
left_limit = 0,
right_limit = N_COLS - 1,
downcount = N_ROWS * N_COLS,
row = 0,
col = 0;
while(printf("%d ", a[row][col]) && --downcount)
if(direction == right)
{
if(++col > right_limit)
{
--col;
direction = down;
++up_limit;
++row;
}
}
else if(direction == down)
{
if(++row > down_limit)
{
--row;
direction = left;
--right_limit;
--col;
}
}
else if(direction == left)
{
if(--col < left_limit)
{
++col;
direction = up;
--down_limit;
--row;
}
}
else /* direction == up */
if(--row < up_limit)
{
++row;
direction = right;
++left_limit;
++col;
}
}
void main()
{
int a[N_ROWS][N_COLS] = {1,2,3,4,5,6,7,8,9,10,11,12,13,14,15};
print_spiral(a);
}
Link for Testing and Download.
Given a matrix of chars, implement a method that prints all characters in the following order: first the outer circle,
then the next one and so on.
public static void printMatrixInSpiral(int[][] mat){
if(mat.length == 0|| mat[0].length == 0){
/* empty matrix */
return;
}
StringBuffer str = new StringBuffer();
int counter = mat.length * mat[0].length;
int startRow = 0;
int endRow = mat.length-1;
int startCol = 0;
int endCol = mat[0].length-1;
boolean moveCol = true;
boolean leftToRight = true;
boolean upDown = true;
while(counter>0){
if(moveCol){
if(leftToRight){
/* printing entire row left to right */
for(int i = startCol; i <= endCol ; i++){
str.append(mat[startRow][i]);
counter--;
}
leftToRight = false;
moveCol = false;
startRow++;
}
else{
/* printing entire row right to left */
for(int i = endCol ; i >= startCol ; i--){
str.append(mat[endRow][i]);
counter--;
}
leftToRight = true;
moveCol = false;
endRow--;
}
}
else
{
if(upDown){
/* printing column up down */
for(int i = startRow ; i <= endRow ; i++){
str.append(mat[i][endCol]);
counter--;
}
upDown = false;
moveCol = true;
endCol--;
}
else
{
/* printing entire col down up */
for(int i = endRow ; i >= startRow ; i--){
str.append(mat[i][startCol]);
counter--;
}
upDown = true;
moveCol = true;
startCol++;
}
}
}
System.out.println(str.toString());
}
Two dimensional N*N Matrix is Square matrix
Idea:
We have to traverse in four different directions to traverse like spiral.
We have to traverse inside matrix once one layer of spiral is over.
So total, we need 5 loops, 4 loops to traverse like spiral and 1 loop to traverse through the layers.
public void printSpiralForm(int[][] a, int length)
{
for( int i = 0 , j = length-1 ; i < j ; i++ , j-- )
{
for( int k = i ; k < j ; k++ )
{
System.out.print( a[i][k] + " " ) ;
}
for( int k = i ; k < j ; k++ )
{
System.out.print(a[k][j] + " ");
}
for( int k = j ; k > i ; k-- )
{
System.out.print(a[j][k] + " ") ;
}
for( int k = j ; k > i ; k-- )
{
System.out.print( a[k][i] + " " ) ;
}
}
if ( length % 2 == 1 )
{
System.out.println( a[ length/2 ][ length/2 ] ) ;
}
}
Just keep it simple -->
public class spiralMatrix {
public static void printMatrix(int[][] matrix, int rows, int col)
{
int rowStart=0;
int rowEnd=rows-1;
int colStart=0;
int colEnd=col-1;
while(colStart<=colEnd && rowStart<=rowEnd)
{
for(int i=colStart;i<colEnd;i++)
System.out.println(matrix[rowStart][i]);
for(int i=rowStart;i<rowEnd;i++)
System.out.println(matrix[i][colEnd]);
for(int i=colEnd;i>colStart;i--)
System.out.println(matrix[rowEnd][i]);
for(int i=rowEnd;i>rowStart;i--)
System.out.println(matrix[i][colStart]);
rowStart++;
colEnd--;
rowEnd--;
colStart++;
}
}
public static void main(String[] args){
int[][] array={{1,2,3,4},{5,6,7,8}};
printMatrix(array,2,4);
}
}
This is my implementation:
public static void printMatrix(int matrix[][], int M, int N){
int level = 0;
int min = (M < N) ? M:N;
System.out.println();
while(level <= min/2){
for(int j = level; j < N - level - 1; j++){
System.out.print(matrix[level][j] + "\t");
}
for(int i = level; i < M - level - 1; i++) {
System.out.print(matrix[i][N - level - 1] + "\t");
}
for(int j = N - level - 1; j > level; j--){
System.out.print(matrix[M - level - 1][j] + "\t");
}
for(int i = M - level - 1; i > level; i-- ){
System.out.print(matrix[i][level] + "\t");
}
level++;
}
}
Here is my solution. Please correct if I'm wrong.
class Spiral:
def spiralOrder(self, A):
result = []
c = []
c.append(A[0])
b = A[1:]
while len(b) > 0:
b = self.rotate(b)
c.append(b[0])
b = b[1:]
for item in c:
for fitem in item:
print fitem,
result.append(fitem)
return result
def rotate(self,a):
b = []
l = zip(*a)
for i in xrange(len(l)-1,-1,-1):
b.append(list(l[i]))
return b
if __name__ == '__main__':
a = [[1, 2, 3,3], [4, 5, 6,6], [7, 8, 9,10]]
s = Spiral()
s.spiralOrder(a)
Slash Top Row -> Transpose -> Flip -> Repeat.
void slashTransposeFlip(int[][] m){
if( m.length * m[0].length == 1){ //only one element left
System.out.print(m[0][0]);
}else{
//print the top row
for(int a:m[0]){System.out.print(a+" ");}
//slash the top row from the matrix.
int[][] n = Arrays.copyOfRange(m,1,m.length);
int[][] temp = n;
int rows = temp.length;
int columns = temp[0].length;
//invert rows and columns and create new array
n = new int[columns][rows];
//transpose
for(int x=0;x<rows;x++)
for(int y=0;y<columns;y++)
n[y][x] = temp[x][y];
//flipping time
for (int i = 0; i < n.length / 2; i++) {
int[] t = n[i];
n[i] = n[n.length - 1 - i];
n[n.length - 1 - i] = t;
}
//recursively call again the reduced matrix.
slashTransposeFlip(n);
}
}
Complexity: Single traverse O(n)
Please let me add my single loop answer with complexity O(n). I have observed that during left-right and right-left traverse of the matrix, there is an increase and decrease by one respectively in the row-major index. Similarly, for the top-bottom and bottom-top traverse there is increase and decrease by n_cols. Thus I made an algorithm for that. For example, given a (3x5) matrix with entries the row-major indexes the print output is: 1,2,3,4,5,10,15,14,13,12,11,6,7,8,9.
------->(+1)
^ 1 2 3 4 5 |
(+n_cols) | 6 7 8 9 10 | (-n_cols)
| 11 12 13 14 15
(-1)<-------
Code solution:
#include <iostream>
using namespace std;
int main() {
// your code goes here
bool leftToRight=true, topToBottom=false, rightToLeft=false, bottomToTop=false;
int idx=0;
int n_rows = 3;
int n_cols = 5;
int cnt_h = n_cols, cnt_v = n_rows, cnt=0;
int iter=1;
for (int i=0; i <= n_rows*n_cols + (n_rows - 1)*(n_cols - 1)/2; i++){
iter++;
if(leftToRight){
if(cnt >= cnt_h){
cnt_h--; cnt=0;
leftToRight = false; topToBottom = true;
//cout << "Iter: "<< iter << " break_leftToRight"<<endl;
}else{
cnt++;
idx++;
//cout << "Iter: "<< iter <<" idx: " << idx << " cnt: "<< cnt << " cnt_h: "<< cnt_h<< endl;
cout<< idx << endl;
}
}else if(topToBottom){
if(cnt >= cnt_v-1){
cnt_v--; cnt=0;
leftToRight = false; topToBottom = false; rightToLeft=true;
//cout << "Iter: "<< iter << " break_topToBottom"<<endl;
}else{
cnt++;
idx+=n_cols;
//cout << "Iter: "<< iter << " idx: " << idx << " cnt: "<< cnt << " cnt_v: "<< cnt_h<< endl;
cout << idx <<endl;
}
}else if(rightToLeft){
if(cnt >= cnt_h){
cnt_h--; cnt=0;
leftToRight = false; topToBottom = false; rightToLeft=false; bottomToTop=true;
//cout << "Iter: "<< iter << " break_rightToLeft"<<endl;
//cout<< idx << endl;
}else{
cnt++;
idx--;
//cout << "Iter: "<< iter << " idx: " << idx << " cnt: "<< cnt << " cnt_h: "<< cnt_h<< endl;
cout << idx <<endl;
}
}else if(bottomToTop){
if(cnt >= cnt_v-1){
cnt_v--; cnt=0;
leftToRight = true; topToBottom = false; rightToLeft=false; bottomToTop=false;
//cout << "Iter: "<< iter << " break_bottomToTop"<<endl;
}else{
cnt++;
idx-=n_cols;
//cout << "Iter: "<< iter << " idx: " << idx << " cnt: "<< cnt << " cnt_v: "<< cnt_h<< endl;
cout<< idx << endl;
}
}
//cout << i << endl;
}
return 0;
}
function spiral(a) {
var s = [];
while (a.length) {
// concat 1st row, push last cols, rotate 180 (reverse inner/outer)...
s = s.concat(a.shift());
a = a
.map(function(v) {
s.push(v.pop());
return v.reverse();
})
.reverse();
}
return s;
}
var arr = [
[1, 2, 3, 4],
[12, 13, 14, 5],
[11, 16, 15, 6],
[10, 9, 8, 7]
];
console.log(spiral(arr));// -> [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16]
arr = [
[0, 1, 2, 3, 4],
[15, 16, 17, 18, 5],
[14, 23, 24, 19, 6],
[13, 22, 21, 20, 7],
[12, 11, 10, 9, 8]
];
console.log(spiral(arr));// -> [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24]
For printing a 2-D matrix consider matrix as a composition of rectangles and/or line where smaller rectangle is fitted into larger one, take boundary of matrix which forms a rectangle to be printed, starting with up-left element each time in each layer; once done with this go inside for next layer of smaller rectangle, in case i don't have a rectangle then it should be line to be printed, a horizontal or vertical. I have pasted the code with an example matrix, HTH.
#include <stdio.h>
int a[2][4] = { 1, 2 ,3, 44,
8, 9 ,4, 55 };
void print(int, int, int, int);
int main() {
int row1, col1, row2, col2;
row1=0;
col1=0;
row2=1;
col2=3;
while(row2>=row1 && col2>=col1)
{
print(row1, col1, row2, col2);
row1++;
col1++;
row2--;
col2--;
}
return 0;
}
void print(int row1, int col1, int row2, int col2) {
int i=row1;
int j=col1;
/* This is when single horizontal line needs to be printed */
if( row1==row2 && col1!=col2) {
for(j=col1; j<=col2; j++)
printf("%d ", a[i][j]);
return;
}
/* This is when single vertical line needs to be printed */
if( col1==col2 && row1!=row2) {
for(i=row1; j<=row2; i++)
printf("%d ", a[i][j]);
return;
}
/* This is reached when there is a rectangle to be printed */
for(j=col1; j<=col2; j++)
printf("%d ", a[i][j]);
for(j=col2,i=row1+1; i<=row2; i++)
printf("%d ", a[i][j]);
for(i=row2,j=col2-1; j>=col1; j--)
printf("%d ", a[i][j]);
for(j=col1,i=row2-1; i>row1; i--)
printf("%d ", a[i][j]);
}
Here is my implementation in Java:
public class SpiralPrint {
static void spiral(int a[][],int x,int y){
//If the x and y co-ordinate collide, break off from the function
if(x==y)
return;
int i;
//Top-left to top-right
for(i=x;i<y;i++)
System.out.println(a[x][i]);
//Top-right to bottom-right
for(i=x+1;i<y;i++)
System.out.println(a[i][y-1]);
//Bottom-right to bottom-left
for(i=y-2;i>=x;i--)
System.out.println(a[y-1][i]);
//Bottom left to top-left
for(i=y-2;i>x;i--)
System.out.println(a[i][x]);
//Recursively call spiral
spiral(a,x+1,y-1);
}
public static void main(String[] args) {
int a[][]={{1,2,3,4},{5,6,7,8},{9,10,11,12},{13,14,15,16}};
spiral(a,0,4);
/*Might be implemented without the 0 on an afterthought, all arrays will start at 0 anyways. The second parameter will be the dimension of the array*/
}
}
//shivi..coding is adictive!!
#include<shiviheaders.h>
#define R 3
#define C 6
using namespace std;
void PrintSpiral(int er,int ec,int arr[R][C])
{
int sr=0,sc=0,i=0;
while(sr<=er && sc<=ec)
{
for(int i=sc;i<=ec;++i)
cout<<arr[sr][i]<<" ";
++sr;
for(int i=sr;i<=er;++i)
cout<<arr[i][ec]<<" ";
ec--;
if(sr<=er)
{
for(int i=ec;i>=sc;--i)
cout<<arr[er][i]<<" ";
er--;
}
if(sc<=ec)
{
for(int i=er;i>=sr;--i)
cout<<arr[i][sc]<<" ";
++sc;
}
}
}
int main()
{
int a[R][C] = { {1, 2, 3, 4, 5, 6},
{7, 8, 9, 10, 11, 12},
{13, 14, 15, 16, 17, 18}
};
PrintSpiral(R-1, C-1, a);
}
int N = Integer.parseInt(args[0]);
// create N-by-N array of integers 1 through N
int[][] a = new int[N][N];
for (int i = 0; i < N; i++)
for (int j = 0; j < N; j++)
a[i][j] = 1 + N*i + j;
// spiral
for (int i = N-1, j = 0; i > 0; i--, j++) {
for (int k = j; k < i; k++) System.out.println(a[j][k]);
for (int k = j; k < i; k++) System.out.println(a[k][i]);
for (int k = i; k > j; k--) System.out.println(a[i][k]);
for (int k = i; k > j; k--) System.out.println(a[k][j]);
}
// special case for middle element if N is odd
if (N % 2 == 1) System.out.println(a[(N-1)/2][(N-1)/2]);
}
}
Java code if anybody is interested.
Input:
4
1 2 3 4
5 6 7 8
9 1 2 3
4 5 6 7
Output: 1 2 3 4 8 3 7 6 5 4 9 5 6 7 2 1
public class ArraySpiralPrinter {
public static void main(String[] args) {
Scanner sc = new Scanner(System.in);
int n = sc.nextInt(); //marrix size
//read array
int[][] ar = new int[n][n];
for (int i = 0; i < n; i++) {
for (int j = 0; j < n; j++) {
ar[i][j] = sc.nextInt();
}
}
printTopRight(0, 0, n - 1, n - 1, ar);
}
//prints top and right layers.
//(x1,y1) to (x1, y2) - top layer & (x1,y2) to (x2, y2)
private static void printTopRight(int x1, int y1, int x2, int y2, int[][] ar) {
//print row values - top
for (int y = y1; y <= y2; y++) {
System.out.printf("%d ", ar[x1][y]);
}
//print column value - right
for (int x = x1 + 1; x <= x2; x++) {
System.out.printf("%d ", ar[x][y2]);
}
//are there any remaining layers
if (x2 - x1 > 0) {
//call printBottemLeft
printBottomLeft(x1 + 1, y1, x2, y2 - 1, ar);
}
}
//prints bottom and left layers in reverse order
//(x2,y2) to (x2, y1) - bottom layer & (x2,y1) to (x1, y1)
private static void printBottomLeft(int x1, int y1, int x2, int y2, int[][] ar) {
//print row values in reverse order - bottom
for (int y = y2; y >= y1; y--) {
System.out.printf("%d ", ar[x2][y]);
}
//print column value in reverse order - left
for (int x = x2-1; x >= x1; x--) {
System.out.printf("%d ", ar[x][y1]);
}
//are there any remaining layers
if (x2 - x1 > 0) {
printTopRight(x1, y1 + 1, x2 - 1, y2, ar);
}
}
}
This is a recursive version in C that I could think of:-
void printspiral (int[][100],int, int, int, int);
int main()
{
int r,c, i, j;
printf ("Enter the dimensions of the matrix");
scanf("%d %d", &r, &c);
int arr[r][100];
int min = (r<c?r:c);
if (min%2 != 0) min = min/2 +1;
for (i = 0;i<r; i++)
for (j = 0; j<c; j++)
scanf ("%d",&arr[i][j]);
printspiral(arr,0,r,c,min );
}
void printspiral (int arr[][100], int i, int j, int k, int min)
{
int a;
for (a = i; a<k;a++)
printf("%d\n", arr[i][a]);
for (a=i+1;a<j;a++)
printf ("%d\n", arr[a][k-1]);
for (a=k-2; a>i-1;a--)
printf("%d\n", arr[j-1][a]);
for (a=j-2; a>i; a--)
printf("%d\n", arr[a][i]);
if (i < min)
printspiral(arr,i+1, j-1,k-1, min);
}
http://www.technicalinterviewquestions.net/2009/03/print-2d-array-matrix-spiral-order.html
here is the best explanation for the above answer :) along with diagram :)
public static void printSpiral1(int array[][],int row,int col){
int rowStart=0,colStart=0,rowEnd=row-1,colEnd=col-1;
int i;
while(rowStart<=rowEnd && colStart<= colEnd){
for(i=colStart;i<=colEnd;i++)
System.out.print(" "+array[rowStart][i]);
for(i=rowStart+1;i<=rowEnd;i++)
System.out.print(" "+array[i][colEnd]);
for(i=colEnd-1;i>=colStart;i--)
System.out.print(" "+array[rowEnd][i]);
for(i=rowEnd-1;i>=rowStart+1;i--)
System.out.print(" "+array[i][colStart]);
rowStart++;
colStart++;
rowEnd--;
colEnd--;
}
}
public class SpiralPrint{
//print the elements of matrix in the spiral order.
//my idea is to use recursive, for each outer loop
public static void printSpiral(int[][] mat, int layer){
int up = layer;
int buttom = mat.length - layer - 1;
int left = layer;
int right = mat[0].length - layer - 1;
if(up > buttom+1 || left > right + 1)
return; // termination condition
//traverse the other frame,
//print up
for(int i = left; i <= right; i ++){
System.out.print( mat[up][i]+ " " );
}
//print right
for(int i = up + 1; i <=buttom; i ++){
System.out.print(mat[i][right] + " ");
}
//print buttom
for(int i = right - 1; i >= left; i --){
System.out.print(mat[buttom][i] + " ");
}
//print left
for(int i = buttom - 1; i > up; i --){
System.out.print(mat[i][left] + " ");
}
//recursive call for the next level
printSpiral(mat, layer + 1);
}
public static void main(String[] args){
int[][] mat = {{1,2,3,4}, {5,6,7,8}, {9,10,11,12}, {13,14,15,16}};
int[][] mat2 = {{1,2,3}, {4,5,6}, {7,8,9}, {10,11,12}};
SpiralPrint.printSpiral(mat2,0);
return;
}
}
Here is my solution in C#:
public static void PrintSpiral(int[][] matrix, int n)
{
if (matrix == null)
{
return;
}
for (int layer = 0; layer < Math.Ceiling(n / 2.0); layer++)
{
var start = layer;
var end = n - layer - 1;
var offset = end - 1;
Console.Write("Layer " + layer + ": ");
// Center case
if (start == end)
{
Console.Write(matrix[start][start]);
}
// Top
for (int i = start; i <= offset; i++)
{
Console.Write(matrix[start][i] + " ");
}
// Right
for (int i = start; i <= offset; i++)
{
Console.Write(matrix[i][end] + " ");
}
// Bottom
for (int i = end; i > start; i--)
{
Console.Write(matrix[end][i] + " ");
}
// Left
for (int i = end; i > start; i--)
{
Console.Write(matrix[i][start] + " ");
}
Console.WriteLine();
}
}
Here's my approach using an Iterator . Note this solves almost the same problem..
Complete code here : https://github.com/rdsr/algorithms/blob/master/src/jvm/misc/FillMatrix.java
import java.util.Iterator;
class Pair {
final int i;
final int j;
Pair(int i, int j) {
this.i = i;
this.j = j;
}
#Override
public String toString() {
return "Pair [i=" + i + ", j=" + j + "]";
}
}
enum Direction {
N, E, S, W;
}
class SpiralIterator implements Iterator<Pair> {
private final int r, c;
int ri, ci;
int cnt;
Direction d; // current direction
int level; // spiral level;
public SpiralIterator(int r, int c) {
this.r = r;
this.c = c;
d = Direction.E;
level = 1;
}
#Override
public boolean hasNext() {
return cnt < r * c;
}
#Override
public Pair next() {
final Pair p = new Pair(ri, ci);
switch (d) {
case E:
if (ci == c - level) {
ri += 1;
d = changeDirection(d);
} else {
ci += 1;
}
break;
case S:
if (ri == r - level) {
ci -= 1;
d = changeDirection(d);
} else {
ri += 1;
}
break;
case W:
if (ci == level - 1) {
ri -= 1;
d = changeDirection(d);
} else {
ci -= 1;
}
break;
case N:
if (ri == level) {
ci += 1;
level += 1;
d = changeDirection(d);
} else {
ri -= 1;
}
break;
}
cnt += 1;
return p;
}
private static Direction changeDirection(Direction d) {
switch (d) {
case E:
return Direction.S;
case S:
return Direction.W;
case W:
return Direction.N;
case N:
return Direction.E;
default:
throw new IllegalStateException();
}
}
#Override
public void remove() {
throw new UnsupportedOperationException();
}
}
public class FillMatrix {
static int[][] fill(int r, int c) {
final int[][] m = new int[r][c];
int i = 1;
final Iterator<Pair> iter = new SpiralIterator(r, c);
while (iter.hasNext()) {
final Pair p = iter.next();
m[p.i][p.j] = i;
i += 1;
}
return m;
}
public static void main(String[] args) {
final int r = 19, c = 19;
final int[][] m = FillMatrix.fill(r, c);
for (int i = 0; i < r; i++) {
for (int j = 0; j < c; j++) {
System.out.print(m[i][j] + " ");
}
System.out.println();
}
}
}
Complete pure C program for any 2D array matrix with given row x column.
#include <stdio.h>
void printspiral(int *p,int r, int c) {
int i=0,j=0,m=1,n=0;
static int firstrun=1,gCol;
if (!p||r<=0||c<=0)
return ;
if(firstrun) {
gCol=c;
firstrun=0;
}
for(i=0,j=0;(0<=i && i<c)&&(0<=j && j<r);i+=m,j+=n) {
printf(" %d",p[i+j*gCol]);
if (i==0 && j==1 && (i+1)!=c) break;
else if (i+1==c && !j) {m=0;n=1;}
else if (i+1==c && j+1==r && j) {n=0;m=-1;}
else if (i==0 && j+1==r && j) {m=0;n=-1;}
}
printspiral(&p[i+j*gCol+1],r-2,c-2);
firstrun=1;
printf("\n");
}
int main() {
int a[3][3]={{0,1,2},{3,4,5},{6,7,8}};
int b[3][4]={{0,1,2,3},{4,5,6,7},{8,9,10,11}};
int c[4][3]={{0,1,2},{3,4,5},{6,7,8},{9,10,11}};
int d[3][1]={{0},{1},{2}};
int e[1][3]={{0,1,2}};
int f[1][1]={{0}};
int g[5][5]={{0,1,2,3,4},{5,6,7,8,9},{10,11,12,13,14},{15,16,17,18,19},{20,21,22,23,24}};
printspiral(a,3,3);
printspiral(b,3,4);
printspiral(c,4,3);
printspiral(d,3,1);
printspiral(e,1,3);
printspiral(f,1,1);
printspiral(g,5,5);
return 0;
}
This question is related to this one: Matrix arrangement issues in php
The answers presented seem to work but are complicated to understand. A very simple way to solve this is divide and conquer i.e., after reading the edge, remove it and the next read will be much simpler. Check out a complete solution in PHP below:
#The source number matrix
$source[0] = array(1, 2, 3, 4);
$source[1] = array(5, 6, 7, 8);
$source[2] = array(9, 10, 11, 12);
$source[3] = array(13, 14, 15, 16);
$source[4] = array(17, 18, 19, 20);
#Get the spiralled numbers
$final_spiral_list = get_spiral_form($source);
print_r($final_spiral_list);
function get_spiral_form($matrix)
{
#Array to hold the final number list
$spiralList = array();
$result = $matrix;
while(count($result) > 0)
{
$resultsFromRead = get_next_number_circle($result, $spiralList);
$result = $resultsFromRead['new_source'];
$spiralList = $resultsFromRead['read_list'];
}
return $spiralList;
}
function get_next_number_circle($matrix, $read)
{
$unreadMatrix = $matrix;
$rowNumber = count($matrix);
$colNumber = count($matrix[0]);
#Check if the array has one row or column
if($rowNumber == 1) $read = array_merge($read, $matrix[0]);
if($colNumber == 1) for($i=0; $i<$rowNumber; $i++) array_push($read, $matrix[$i][0]);
#Check if array has 2 rows or columns
if($rowNumber == 2 || ($rowNumber == 2 && $colNumber == 2))
{
$read = array_merge($read, $matrix[0], array_reverse($matrix[1]));
}
if($colNumber == 2 && $rowNumber != 2)
{
#First read left to right for the first row
$read = array_merge($read, $matrix[0]);
#Then read down on right column
for($i=1; $i<$rowNumber; $i++) array_push($read, $matrix[$i][1]);
#..and up on left column
for($i=($rowNumber-1); $i>0; $i--) array_push($read, $matrix[$i][0]);
}
#If more than 2 rows or columns, pick up all the edge values by spiraling around the matrix
if($rowNumber > 2 && $colNumber > 2)
{
#Move left to right
for($i=0; $i<$colNumber; $i++) array_push($read, $matrix[0][$i]);
#Move top to bottom
for($i=1; $i<$rowNumber; $i++) array_push($read, $matrix[$i][$colNumber-1]);
#Move right to left
for($i=($colNumber-2); $i>-1; $i--) array_push($read, $matrix[$rowNumber-1][$i]);
#Move bottom to top
for($i=($rowNumber-2); $i>0; $i--) array_push($read, $matrix[$i][0]);
}
#Now remove these edge read values to create a new reduced matrix for the next read
$unreadMatrix = remove_top_row($unreadMatrix);
$unreadMatrix = remove_right_column($unreadMatrix);
$unreadMatrix = remove_bottom_row($unreadMatrix);
$unreadMatrix = remove_left_column($unreadMatrix);
return array('new_source'=>$unreadMatrix, 'read_list'=>$read);
}
function remove_top_row($matrix)
{
$removedRow = array_shift($matrix);
return $matrix;
}
function remove_right_column($matrix)
{
$neededCols = count($matrix[0]) - 1;
$finalMatrix = array();
for($i=0; $i<count($matrix); $i++) $finalMatrix[$i] = array_slice($matrix[$i], 0, $neededCols);
return $finalMatrix;
}
function remove_bottom_row($matrix)
{
unset($matrix[count($matrix)-1]);
return $matrix;
}
function remove_left_column($matrix)
{
$neededCols = count($matrix[0]) - 1;
$finalMatrix = array();
for($i=0; $i<count($matrix); $i++) $finalMatrix[$i] = array_slice($matrix[$i], 1, $neededCols);
return $finalMatrix;
}
// Program to print a matrix in spiral order
#include <stdio.h>
int main(void) {
// your code goes here
int m,n,i,j,k=1,c1,c2,r1,r2;;
scanf("%d %d",&m,&n);
int a[m][n];
for(i=0;i<m;i++)
{
for(j=0;j<n;j++)
{
scanf("%d",&a[i][j]);
}
}
r1=0;
r2=m-1;
c1=0;
c2=n-1;
while(k<=m*n)
{
for(i=c1;i<=c2;i++)
{
k++;
printf("%d ",a[r1][i]);
}
for(j=r1+1;j<=r2;j++)
{
k++;
printf("%d ",a[j][c2]);
}
for(i=c2-1;i>=c1;i--)
{
k++;
printf("%d ",a[r2][i]);
}
for(j=r2-1;j>=r1+1;j--)
{
k++;
printf("%d ",a[j][c1]);
}
c1++;
c2--;
r1++;
r2--;
}
return 0;
}

Resources