Ideas, Knowledge, Technology, Computer Science, Experience associated with my work and some geeky stuff I progressively encounter during my journey towards enlightenment. Read on…

  • RSS RSS Feed

    • Cloud Computing
      It’s been really long, since I last wrote a tech post. In this post, I’m just sharing few useful links to get started on Cloud Computing, whether you’re a developer, quality engineer, business leader, or a project manager intending to get started with Cloud Computing. I’m currently designing systems and services, for a platform that’s […]
    • The Pragmatic Programmer
      I finished reading The Pragmatic Programmer by Andrew Hunt and David Thomas. It’s not a new book in the market but I was curious to read this. The technology topics covered, are not any different from those found in most software engineering books, but the way they’re presented using Pragmatic Philosophy Approach, is remarkable. Code […]
    • 2013 in review
      The stats helper monkeys prepared a 2013 annual report for this blog. Here’s an excerpt: A San Francisco cable car holds 60 people. This blog was viewed about 1,200 times in 2013. If it were a cable car, it would take about 20 trips to carry that many people. Click here to see the […]
    • Goodbye, Ness!
      It had to happen sometime. I thought Feb 2013 was the right time. I quit Ness after a long 5 years and 4 months of stay, in Feb. I joined FICO (formerly, Fair Isaac) last Feb.  While I get an opportunity to work with many varied stakeholders like Scientists, Architect, Product Management, Peer Developers, PMO, Technical Publications and also […]
    • Meta: information retrieval from Tweets
      I pick significant problems randomly sometimes and enjoy solving them, or at least attempt designing api :-). Here’s one such problem! Problem: How’d you go about finding meta information about a person’s tweets? NOTE: a) Tweet == Twitter updates b) Meta information –> Loosely defined. You can interpret it anyway you want –> Frequency, topics, follower […]
  • Twitter Updates

Archive for July, 2011

Find the kth Smallest Element in Array in Linear Time

Posted by sanstechbytes on July 24, 2011

Finding the kth smallest element in array is an application related to Order Statistics problem. For details on the algo used here, watch the video lecture embedded below.

Robert Sedgewick, whose lecture this approach has been inspired from, proposes random shuffle mechanism and partitioning strategy to select the kth smallest element. The strategy requires maintaining variables lo and hi to delimit the subarray that contains the index k of the element to be selected, using partitioning to shrink the size of the subarray (Step 2). The assumption made is that the input array to the algo doesn’t contain duplicates (This can be carried out as a preprocessing step, which is performed as first step in the algo below, though). Runs in linear time, O(n).

1) Preprocess array to remove duplicates.
2) Rearrange the array from a[lo] through a[hi] and return an integer j such that a[lo] through a[j-1] are less than a[j], and a[j+1] through a[hi] are greater than a[j].
a) If k is equal to j, done.
b) Else If k is less than j, then continue working in the right subarray (by changing the value of hi to j-1).
c) Else If k is greater than j, then continue working in the right subarray (by changing lo to j+1).
3) Repeat step (2), until the array consists just of k.

MIT Video Lecture on Order Statistics:

Java Code:

/** Sedgewick's partitioning approach */
 public class QuickSelect {

private static Random random = new Random();

public static Comparable select(Comparable[] a, int k) {
 if (k = a.length) {
 throw new RuntimeException("Selected element out of bounds");


int lo = 0, hi = a.length - 1;
 while (hi > lo) {
 int i = partition(a, lo, hi);
 if (i > k)
 hi = i - 1;
 else if (i < k)
 lo = i + 1;
 return a[i];
 return a[lo];

private static int partition(Comparable[] a, int lo, int hi) {
 int i = lo;
 int j = hi + 1;
 Comparable v = a[lo];
 while (true) {
 // find item on lo to swap
 while (less(a[++i], v))
 if (i == hi)
 // find item on hi to swap
 while (less(v, a[--j]))
 if (j == lo)
 break; // redundant since a[lo] acts as sentinel
 // check if pointers cross
 if (i >= j)
 exch(a, i, j);
 // put v = a[j] into position
 exch(a, lo, j);
 // with a[lo .. j-1] <= a[j] <= a[j+1 .. hi]
 return j;

public static void shuffle(Object[] a) {
 int n = a.length;
 for (int i = 0; i < n; i++) {
 int r = i + uniform(n - i); // between i and N-1
 Object temp = a[i];
 a[i] = a[r];
 a[r] = temp;

 * Return an integer uniformly between 0 and N-1.
 public static int uniform(int N) {
 return random.nextInt(N);

// is v < w ?
 private static boolean less(Comparable v, Comparable w) {
 return (v.compareTo(w) < 0);

// exchange a[i] and a[j]
 private static void exch(Object[] a, int i, int j) {
 Object swap = a[i];
 a[i] = a[j];
 a[j] = swap;

// Read strings from standard input, sort them, and print.
 public static void main(String[] args) {
 // Read input from StdIn here
 // String[] a = new String[]{"aa", "aaa", "bb", "ab", "ca", "bc",
 // "cab", "cc", "cad" };
 Integer[] a = new Integer[] { 2, 11, 10, 6, 8, 99, 7, 100, 101, 28,
 9 };
 int k = 2;
 Object kth =, k - 1); // returns a[k-1]
 System.out.println(k + "th smallest element in given array = "
 + kth);


Posted in Algorithm | Tagged: | Leave a Comment »

Pythagorean Triples

Posted by sanstechbytes on July 10, 2011

A quick background on properties of Pythagorean triples can be got from here and here.

When I was browsing through topics in algorithm forum on , I found , sort of, an interesting challenge in finding Pythagorean Triples in the given array, saying:  Can’t be achieved in no better than O(n*n) time complexity(a unanimous opinion). This enthused me to attempt writing a faster than O(n*n) algo, but after much analysis and code completion, I couldn’t help joining the participants. In below lines, the code snippet follows. I wonder, if any algorithm enthusiast or researcher could point me to a faster algorithm than O(n*n), if there exists any.

Follow comments in-line to help yourself understand the algorithm used. This algorithm achieves O(k*n*n) time complexity where k is nearly equal to n/3  in most cases. For brevity, it’s an O(n*n) algo.

// O(n+n*log n+k*n*n) = O(k*n*n)= O(n*n) algo where k is nearly n/3 in
// most cases
// Returns the number of triples found.
// Returns
// -1, when n < 3,
// 1 when n = 3 and/or a triple is found and
// 0 when no triple found
public static int pythtriples(int[] pyth) {
int n = pyth.length;

// No duplicate removal algo is used before sort, to avoid the cost of
// O(n),as this will be taken care of below,in for loop.
// (Trivial) Shifting mechanism should be used in
// array with new length in-place(trivial)in case,
// either the element occurring in multiple places is 0 or neg integer

// Square all the numbers – O(n) time complexity.
for (int i = 0; i < n; i++) {
pyth[i] = pyth[i] * pyth[i];

// Sort them – O(n * log n) time complexity

if (n < 3) {
return -1; // error : No of elements must be >= 3

// Simple case
if (n == 3) {
if (pyth[2] == pyth[1] + pyth[0])
return 1;
return 0;

// Compare pair-wise: if(a[i] == (a[j] + a[k]) then, it’s a triple),
// k<i, j<i
// O(k* n * n) time complexity, where k is nearly n/3, can be
int triples = 0, foundtripleLarge = 0;
for (int i = 2; i < n; i++) {
for (int j = 0; j < i; j++) {
for (int k = i – 1; k >= 0; k–) {
// If already found, don’t compare again for 3SUM. Avoids
// unnecessary check for duplicates
if (foundtripleLarge == pyth[i]) {
if (pyth[i] == (pyth[j] + pyth[k])) {
System.out.println(“triples found: ” + pyth[i] + ” = ”
+ pyth[j] + ” + ” + pyth[k]);
triples++; // Counter for triples
foundtripleLarge = pyth[i];
return triples;

Math: The pair-wise comparison follows: 1^2 + 2^2 + 3^2 + …..+n^2 = n(n+1)(2n+1)/6, nearly equal to 3n^2, same as our O(kn^2) algo above!

Posted in Algorithm | Tagged: | Leave a Comment »

Reliable Processing of HTTP requests

Posted by sanstechbytes on July 6, 2011

I thought of sharing an interesting update from my circle of friends. Keerthidhar Dongre, an ex-colleague and a friend of mine, had filed a patent in December of 2008, on his claims about Reliable Processing of HTTP requests.
His patent has been accepted by USPTO.The full-text for this patent US007975047  can be read here. An amazing feat, indeed!

Posted in Research | Leave a Comment »