So here is a programming question once again,
Consider a list of unsorted positive numbers in which each number has a pair except one of the number, the problem is to find that number in minimum complexity with O(1) extra space. The complexity in this case should not be more than O(n).
For instance, in the list {5,7,4,9,5,7,4} the answer should be 9. I will post my solution in the comments some time later.
The solution is as follows remember the list is not sorted,
FACT 1: XORing a number with 0 will return the number itself.
FACT 2: XORing a number with itself will return 0.
So when we XOR all the elements with each other in a list of number in which each number is repeated once except one number we will get the non repeated number. For example if List is {5,4,6,4,5} we will get 6 if we execute this: 5 XOR 4 XOR 6 XOR 4 XOR 5.
Here is the C Sharp Source to do the same.
int Foo(int [] nList)
{
int result = 0;
for (int i = 0; i < nList.Length; i++)
result ^= nList[i];
return result;
}
In C Sharp ‘^’ is the operator for XOR, This piece of code takes O(n) time and O(1) extra space.
By: Haroon Saeed on March 12, 2008
at 7:44 am
I think a little bit modification require …
result ^= nList[i]
By: Muhammad Qasim Pasta on March 12, 2008
at 7:13 pm
and what about make it little bit more efficient …
int Foo(int [] nList)
{
int result = nList[0]; // we know the first element
for (int i = 1; i < nList.Length; i++) // running one less than total numbers
result ^= nList[i];
return result;
}
By: Muhammad Qasim Pasta on March 12, 2008
at 7:16 pm
Ohh well Qasim Pasta!! there was a typo, ammended it..
, well I dont think capturing the first element and running the loop n – 1 times will make any difference, though your comment and intention noted 
…
Thx for poiting out the typo..
By: Haroon Saeed on March 12, 2008
at 7:51 pm
ok, I have another solution
Fact: all of list elements has another element equivalent except for one element, which means that list length is an odd number
if you start from the beginning the list and add the numbers to each other, until you reach the middle number, then you start subtracting numbers from the total until you are done, then the final count will be the number with no corresponding peer
int Foo(int[] arr)
{
int count = arr[0];
for(int i = 0; i < arr.Length; i++)
{
if( i <= (size – 1) / 2)
count += arr[i];
else
count -= arr[i];
}
return count;
}
By: Hisham abd El-Hafez on April 2, 2008
at 4:31 pm
Create Map
Loop thru your array and check
if elememt in aaray exist in map , if not then add
if exists then remove
Finally only one element will left will be your non repeated elemet.
By: san on April 5, 2008
at 5:57 am
san, the problem definition states that you should solve it in O(1) extra space, using your map, the extra space complexity will be O(n-1)
By: Hisham abd El-Hafez on April 6, 2008
at 7:43 am
@ Hisham abd El-Hafez:
Actually your solution works only if the pairs are quasi symmetrically placed and if your pairless number is not the center number then you end up performing the same operation on both the numbers of atlease 1 pair;
eg: 4 4 5 7 5
according to your solution we actuallly have
4 + 4 + 5 – 7 – 5 = 1
By: Vinayak on April 30, 2008
at 1:21 pm
Vinayak,
Yeah you are right, my solution will only work if the array is symmetric.
my mistake
By: Hisham abd El-Hafez on May 12, 2008
at 1:30 pm
By the way.. Let’s manually run the proposed algorithm for the following list: ( 4,5,4,7,5 )
(4 ^ 5) ^ 4 ^ 7 ^ 5
( 0 ^ 4 ) ^ 7 ^ 5
(4 ^ 7 ) ^ 5
0 ^ 5 = 5 .. and 5 is not the non repeating element….
By: Arthur Rodrigues on August 2, 2008
at 1:24 pm
@Arthur Rodrigues:
(4 ^ 5) = 1, and it works
By: yashez on September 10, 2008
at 2:04 am
The proposed algorithm clearly does not work.
if the list is [5, 4, 6, 4, 5]
then,
result = (((5 XOR 4) XOR 6) XOR 4) XOR 5
= (( 0 XOR 6) XOR 4) XOR 5
= (6 XOR 4) XOR 5
= 0 XOR 5
= 5
Perhaps I’ve misunderstood ?
There’s an intuitive solution to this problem, based on quick sort. It’s O (n log n) though
Pick an element at random, the pivot, and then move the elements less than this number into one set and the elements greater into another set.
If your two sets are of equal size, then stop and return the pivot.
Repeat on the list that has an odd size.
By: Owen on September 12, 2008
at 1:08 am
Sorry
scratch that, I haven’t had my morning coffee
Just sort O (n log n) then scan through (O (n))
By: Owen on September 12, 2008
at 1:10 am
Actually,
My original algorithm will work and will be slightly faster than quick sort because you only recurse down a fraction of the tree.
But, it’s still not linear time.
By: Owen on September 12, 2008
at 1:13 am