[NumPy-Tickets] [NumPy] #1488: vectorize(): could the dtype of the output array be the "smallest possible dtype"?

NumPy Trac numpy-tickets@scipy....
Fri May 21 09:24:38 CDT 2010


#1488: vectorize(): could the dtype of the output array be the "smallest possible
dtype"?
------------------------------+---------------------------------------------
 Reporter:  lebigot           |       Owner:  somebody
     Type:  enhancement       |      Status:  new     
 Priority:  normal            |   Milestone:  2.0.0   
Component:  Other             |     Version:  1.3.0   
 Keywords:  vectorize, dtype  |  
------------------------------+---------------------------------------------
 It would be nice if vectorize()'d functions returned arrays with the
 "smallest possible dtype".  Maybe such a feature could be added as a flag
 in vectorize's arguments?

 Thus, the result of a vectorized function would be more robust; currently,
 if the first call to the vectorized function returned an array of
 integers, all its subsequent calls return integers even when the user
 really expects floats.  This problem might be hard to detect, in a long
 and complicated matrix calculation.

 This could be one solution to the ticket in
 http://projects.scipy.org/numpy/ticket/1487

-- 
Ticket URL: <http://projects.scipy.org/numpy/ticket/1488>
NumPy <http://projects.scipy.org/numpy>
My example project


More information about the NumPy-Tickets mailing list