Robert, Thanks for solving that puzzle! I'll get our group on the same 1.0.x numpy release.
Take care, James ************************** Harvard University Dept. of Astronomy 60 Garden Street MS-10 Cambridge, MA 02138 phone 617.496.5988 lab 617.495.3267 email [EMAIL PROTECTED] web http://www.cfa.harvard.edu/~jbattat ************************** On Wed, 23 Jan 2008, Robert Kern wrote: > James Battat wrote: > > Hi, > > > > numpy.where() returns different things on my windowsXP machine than on my > > collaborator's mac osx machine. > > > > For example, consider: > > >>> import numpy > > >>> a = numpy.array([1,2,3,4]) > > >>> b = numpy.where( a > 2 ) > > > > On WindowsXP (python 2.4.2, numpy 1.0.1), I get: > > >>> print b > > (array([2, 3]), ) > > >>> print b[0] > > [2 3] > > > > While in OSX (python 2.4.3, numpy 0.9.6), he gets: > > >>> print b > > [2, 3] > > >>> print b[0] > > 2 > > > > This matters because we use where() to get indices to loop over. On the > > mac you can then loop through the b array: > > > > for item in b: > > if b > 5: > > print 'hi' > > > > But this approach fails on a pc because: > > if b > 5: > > >>> ValueError: The truth value of an array with more than one element > > >>> is ambiguous. Use a.any() or a.all() > > > > > > I'd like to avoid having to do: > > if mac: > > b = numpy.where( a > 5) > > if pc: > > b = numpy.where( a > 5)[0] > > > > Has anybody else notice dealt with this? Is this a mac/pc difference or a > > numpy 1.01 vs. numpy 0.9.6 difference? > > The latter. Before the 1.0 release, we tried to make it clear that we were > still > experimenting with the APIs and that they might change up until 1.0 got > released. > > -- > Robert Kern > > "I have come to believe that the whole world is an enigma, a harmless enigma > that is made terrible by our own mad attempt to interpret it as though it > had > an underlying truth." > -- Umberto Eco > _______________________________________________ > Numpy-discussion mailing list > [email protected] > http://projects.scipy.org/mailman/listinfo/numpy-discussion > _______________________________________________ Numpy-discussion mailing list [email protected] http://projects.scipy.org/mailman/listinfo/numpy-discussion
