ie key -----> A*Atrans--------->after multiplication the result will
be a 2D array which is declared as double (matrix) lets say the result
be Matrix "Ekey"(double[][] Ekey)
value ------> Atrans*D ---------> after multiplication the result
will be Matrix "Eval" (double[][] Eval).
After tat i need to emit these matrix to reducer for further calculations.
so in mapper
context.write(Ekey,Eval);
reducer
i need to do further calculations with these Ekey nd Eval.ie
calculate (I/v + Ekey)*Eval. where I/v is a matrix with order of Ekey
public int compareTo(MatrixWritable o) {
// TODO Auto-generated method stub
boolean ret = Arrays.equals(values, o.values);
}
I am really confused what to check, as i dnt get a complete idea of
compareTo() while relating it to mapreduce.
my input will be a file wit numerical data say 'A' . from that i am
calculating Ekey and Eval in mapper.
context.write(new MatrixWritable(Ekey),new
MatrixWritable(Eval));
so what to compare. whether keys?if same key is coming...some thing
like that or?
On Mon, Nov 4, 2013 at 5:07 PM, Mirko Kämpf <[email protected]> wrote:
> You just have to implement the WritableComparable interface. This might be
> straight forward for any simple data type. Have look on:
>
> public int compareTo(MyWritableComparable o) {
> int thisValue = this.value
>
> int thatValue = o.value;
>
> return (thisValue < thatValue ? -1 : (thisValue==thatValue ? 0 : 1));
> ...
>
> But a matrix is something different.
>
> Now its your turn. Find out how to order your matrix objects. This can be any
> type of order which makes sense. Maybe the total sum or the total average of
> all values or whatever. It has to calc only -1, 0 or 1, depending on the
> given matrixes. I would assume that 0 will be given back if all pairs of
> elements from both matrizes with the same indexes are equal.
>
> But why do you need the Matrix one as the KEY in your reducer?
>
> Maybe it is better to use Custom object only as the Value, together with an
> arbitrary Key, which is WritableComparable ... but this depends on, what you
> want to calc in your reducer.
>
> If you share the description of the problem and the algorithm which solves
> it, we will find out whats the best way.
>
> What does you algorithm calculate?
>
>
> Best wishes
>
> Mirko
>
>
>
>
> 2013/11/4 unmesha sreeveni <[email protected]>
>
>> I am trying to emit a matrix(2D double array) as key and another matrix
>> as value.
>>
>> ie context.write(new myclass(matrixa),new myclass(matrixb));
>> * so how to compare them
>>
>>
>> On Mon, Nov 4, 2013 at 4:16 PM, unmesha sreeveni
>> <[email protected]>wrote:
>>
>>> Thanks Dieter De Witte
>>>
>>>
>>> On Mon, Nov 4, 2013 at 4:11 PM, Dieter De Witte <[email protected]>wrote:
>>>
>>>> A writableComparable is an interface. A key in mapreduce needs to
>>>> implement it, since the keys need to be compared for equality. So it has
>>>> nothing to do with key value pairs as such.
>>>>
>>>> Regards, Dieter
>>>>
>>>>
>>>> 2013/11/4 unmesha sreeveni <[email protected]>
>>>>
>>>>>
>>>>> what is CompateTo() in WritableComparable?
>>>>> Is that comparing a key and value or comparing key and key,value and
>>>>> value
>>>>>
>>>>> --
>>>>> *Thanks & Regards*
>>>>>
>>>>> Unmesha Sreeveni U.B
>>>>>
>>>>> *Junior Developer*
>>>>>
>>>>> *Amrita Center For Cyber Security *
>>>>>
>>>>>
>>>>> * Amritapuri.www.amrita.edu/cyber/ <http://www.amrita.edu/cyber/>*
>>>>>
>>>>
>>>>
>>>
>>>
>>> --
>>> *Thanks & Regards*
>>>
>>> Unmesha Sreeveni U.B
>>>
>>> *Junior Developer*
>>>
>>> *Amrita Center For Cyber Security *
>>>
>>>
>>> * Amritapuri.www.amrita.edu/cyber/ <http://www.amrita.edu/cyber/>*
>>>
>>
>>
>>
>> --
>> *Thanks & Regards*
>>
>> Unmesha Sreeveni U.B
>>
>> *Junior Developer*
>>
>> *Amrita Center For Cyber Security *
>>
>>
>> * Amritapuri.www.amrita.edu/cyber/ <http://www.amrita.edu/cyber/>*
>>
>
>
--
*Thanks & Regards*
Unmesha Sreeveni U.B
*Junior Developer*
*Amrita Center For Cyber Security*
* Amritapuri.www.amrita.edu/cyber/ <http://www.amrita.edu/cyber/>*
