[NumPy-Tickets] [NumPy] #1949: Upcasting a float32 array to complex64 instead of complex128 when adding a complex128 scalar

[NumPy Trac]

#1949: Upcasting a float32 array to complex64 instead of complex128 when adding a
complex128 scalar
---------------------+------------------------------------------------------
 Reporter:  tihocan  |       Owner:  somebody   
     Type:  defect   |      Status:  new        
 Priority:  normal   |   Milestone:  Unscheduled
Component:  Other    |     Version:  1.6.0      
 Keywords:           |  
---------------------+------------------------------------------------------
 This is with numpy 1.6.0 and 1.6.1 under Linux x86_64, testing the upcast
 mechanism of "scalar + array":

 {{{
 >>> import numpy; print (numpy.array(2, dtype=numpy.complex128) +
 numpy.ones(3, dtype=numpy.float32)).dtype
 complex64
 }}}

 Since it has to upcast my array (float32 is not "compatible enough" with
 complex128), why does it upcast it to complex64 instead of complex128?
 As far as I can tell 1.4.x and 1.5.x versions of numpy are indeed
 upcasting to complex128.

 Also note that:

    * When replacing "numpy.ones" with "numpy.array" it yields complex128
 (expected upcast of scalar addition of complex128 with float32)
    * The behavior is similar if instead of "2" I use a number which cannot
 be represented exactly with a complex64 (so it's not a rule about picking
 the smallest data type able to exactly represent the result)

 This issue was reported on the mailing list in:
 http://mail.scipy.org/pipermail/numpy-discussion/2011-August/057981.html

-- 
Ticket URL: <http://projects.scipy.org/numpy/ticket/1949>
NumPy <http://projects.scipy.org/numpy>
My example project