[SciPy-user] Replacing elements in matrix

steve schmerler elcorto at gmx.net
Tue Apr 26 13:07:34 CDT 2005



David M. Cooke wrote:
> steve schmerler <elcorto at gmx.net> writes:
> 
> 
>>Dimitri D'Or wrote:
>>
>>>Hello,  I come from the Matlab world and I'm used to write
>>>operations such as: A[indexi,indexj]=A[indexi,indexj]+B where A is a
>>>n by n matrix, indexi is i by 1 vector, indexj is a j by 1 vector
>>>and B is a i by j matrix.  With this operation, some of the elements
>>>of A are replaced by their original value augmented by some value
>>>coming from B.  for example, take
>>>A=array([[1,5,7,8,2],[5,3,4,6,7],[9,4,6,7,2],[0,2,0,3,4]])
>>>indexi=array([0,2]) indexj=array([1,2,4])
>>>B=array([[10,20,30],[100,200,300]])  With this, Matlab results would
>>>be A[indexi,indexj]=array([[5, 7, 2],[4, 6, 2]]) and
>>>A[indexi,indexj]=A[indexi,indexj]+B yields
>>>A=array([[1,15,27,8,32],[5,3,4,6,7],[9,104,206,7,302],[0,2,0,3,4]])
>>>I would like to make the same operation with Python. Would you
>>>please propose me a compact code for achieving that?
>>
>>Hmm, I'm affraid this isn't straightforward. To select arbitrary rows and cols from A you can use
>>Numeric's take()
>>
>>	In [10]: take(take(A,(0,2),axis=0),(1,2,4),axis=1)
>>	Out[10]:
>>	array([[5, 7, 2],
>>        	       [4, 6, 2]])
>>
>>but any operation on the selected submatrix won't alter A.
> 
> 
>>With slicing, e.g. select the upper 2x2 block
>>
>>	In [14]: A[0:2,0:2]
>>	Out[14]:
>>	array([[1, 5],
>>                [5, 3]])
>>
>>you also can't alter A by performing some operation on the subblock, e.g.
> 
> 
> eh? No, a slice is a view of the matrix, so it also operates on the
> original matrix.
> 
> 
>>	A[0:2,0:2]*100
>>
>>doesn't change A.
> 
> 
> Of course, that doesn't change an array at all (you'll get a new 2x2
> matrix). This will:
> 
> A[0:2,0:2] *= 100
> 
> That'll change the submatrix A[0:2,0:2], and A.
> 

Sorry, of course you're right :)

> 
>>Btw, I don't know how to select _arbitrary_ rows and cols with slicing (hints?).
> 
> 
> You mean, say, the i'th row? That's A[i]. 

Sure.

> If you need an arbitrary
> selection of rows, you're out of luck as far as making one matrix
> that's a view of A (but you can make a copy as you did above with take()).
> 

Playing arround with take and put would do it but I don't see a simple MATLAB-like solution with slicing, though.

Cheers,
Steve

-- 
Man is the best computer we can put aboard a spacecraft ... and the only one that can be
mass produced with unskilled labor.  - Wernher von Braun



More information about the SciPy-user mailing list