OLOLO - Onotole needs your help


Onotole has a lot of pyani. Each pyani has a number, writing on it. Pyanis with equal numbers are indistinguishable. Onotole knows everything, so, he knows that each pyani appeared twice, and only one pyani is unique. He wants to get вздръжни эффект, and he needs the unique pyani. Given the list of pyanis denote which one of them appeared once (it is guaranteed that other pyanis appeared twice).

Input

First line of input contains number of pyanis N <= 500 000. Next N lines contain a single positive integer 1 <= Pi <= 10^9.

Output

Output one positive integer on pyani, which appeared once.

Example

Input:
3
1
8
1

Output: 8

Onotole has found not optimal AC algorithms, so all solutions will be rejudged. He is watching you.


hide comments
.:: Jarv1s ::.: 2015-03-15 19:06:30

Learned a new thing!

Francky: 2015-02-03 19:08:28

The issue seems solved.

Mitch Schwartz: 2015-02-01 14:58:34

I downloaded one of my previously uploaded images and it differed from the original by one byte: the one on SPOJ server has a 0x0A byte appended to the beginning.

Francky → Nice catch!
Independently, could you confirm that RSS is broken? (I've yet email admins for that too).

(Mitch) I just thought about that before reading your reply. It looks like it's actually the same issue: an extra '\n' appearing at the beginning of the file. You can see all the RSS data is still there if you view source. And this '\n' is also present for problem description pages. I don't know if the original files have all been modified, or if they are just being generated on the fly with this extra byte.

Francky → Nice combo catch!! Then it sounds not so independent ;-)
Let's wait for admin's patch. There should have been some tests on their side ; I think your option 2 is the good one (on the fly). And all things should be cleared in the same time.

Last edit: 2015-02-01 16:51:07
Scarface: 2015-01-31 22:22:53

Image isn't visible.

--Francky--> It seems it's a general problem not only for this one. Email to admins.

Last edit: 2015-01-31 22:40:18
Devashish Mathur: 2015-01-27 15:16:45

Great Problem... Learnt something new...

agaurav77: 2014-12-28 13:46:36

Cool question, giving a hint would spoil it. But believe me, doing this one is worth it.

Madhur Bhargava: 2014-10-14 05:31:03

Java Guys: Fastest input handling will be required for this!

numerix: 2013-04-15 16:07:15

I don't know what the problemsetter considers to be a "not optimal algorithm", but anyway it isn't fair to reduce timelimit after several AC submissions (e.g. my Python solution). Decision about timelimit should be made BEFORE a problemset is published.


Added by:Efim
Date:2010-11-04
Time limit:0.134s
Source limit:50000B
Memory limit:1536MB
Cluster: Cube (Intel G860)
Languages:All except: ASM64