Ohh my bad. the complexity should be
O(nlogn) + O(mlogm) +O(m) = O(tlogt) where t=max(m,n)

On Sat, Sep 3, 2011 at 11:46 PM, mohit verma <[email protected]> wrote:

> create a binary search tree by scanning the whole array and if the value
> already exists increase count field in that node O(nlogn). Now traverse the
> tree in any order by creating another tree with kery - count. O(nlogn).
> Doing reverse inorder traversal print value field of each node count number
> of times O(n).
>
> overall complexity - O(nlogn)+O(nlogn)+O(n) = O(nlogn).
>
>
> On Sat, Sep 3, 2011 at 11:16 PM, Siddhartha Banerjee <
> [email protected]> wrote:
>
>> perhaps we can make it O(mlogm), m= no of distinct elements... just create
>> a hash table and store the count of the number of time elements occur...
>> O(n), now sort the m elements and proceed as above...
>> it is obviously not possible to do it faster than O(mlogm), where m = no
>> of distinct elements...
>> so order = O(n+mlogm)=O(mlogm)(???, assuming m!<<n)
>>
>>  --
>> You received this message because you are subscribed to the Google Groups
>> "Algorithm Geeks" group.
>> To post to this group, send email to [email protected].
>> To unsubscribe from this group, send email to
>> [email protected].
>> For more options, visit this group at
>> http://groups.google.com/group/algogeeks?hl=en.
>>
>
>
>
> --
> ........................
> *MOHIT VERMA*
>
>


-- 
........................
*MOHIT VERMA*

-- 
You received this message because you are subscribed to the Google Groups 
"Algorithm Geeks" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/algogeeks?hl=en.

Reply via email to