python - Numpy vectorize, using lists as arguments -

the numpy vectorize function useful, doesn't behave when function arguments lists rather scalars. example:

import numpy np  def f(x, a):     print "type(a)=%s, a=%s"%(type(a),a)     return sum(a)/x  x = np.linspace(1,2,10) p = [1,2,3]  f2 = np.vectorize(f)  f(x,p) f2(x,p) 

gives:

type(a)=<type 'list'>, a=[1, 2, 3] type(a)=<type 'numpy.int64'>, a=1  traceback (most recent call last):   file "vectorize.py", line 14, in <module>     f2(x,p)   file "/usr/local/lib/python2.6/dist-packages/numpy/lib/function_base.py", line 1824, in __call__     theout = self.thefunc(*newargs)   file "vectorize.py", line 5, in f     return sum(a)/x typeerror: 'numpy.int64' object not iterable 

i understand function f works just fine without vectorizeing it, i'd know how (in general) vectorize function arguments take in lists rather scalar.

your question doesn't make clear precisely output see vectorized function, i'm going assume same list (a) applied argument every invocation of f() (ie once each element in x array)

the vectorized version of function ensures arguments arrays, , applies numpy's broadcasting rules determine how these arguments should combined.

as np.array's wrapping of np.ndarray, coercion of arguments arrays tries helpful, automatically converting list array containing same elements, rather making array dtype=object contains list sole element. of time want, in case, "smart" behaviour coming bite you.

while there may way instruct numpy treat inputs vectors, there 2 straightforward ways behaviour you're after:

1. manually create array dtype=object work within broadcasting rules
2. curry value prior vectorizing function

1. dtype=object

numpy arrays derive efficiency storing 1 type of item, can still contain arbitrary python objects specifying stored data type python objects:

list_obj_array = np.ndarray((1,), dtype=object) list_obj_array[0] = [1,2,3] f2(x,list_obj_array)  # using definition above 

prints:

type(a)=<type 'list'>, a=[1, 2, 3] type(a)=<type 'list'>, a=[1, 2, 3] type(a)=<type 'list'>, a=[1, 2, 3] type(a)=<type 'list'>, a=[1, 2, 3] type(a)=<type 'list'>, a=[1, 2, 3] type(a)=<type 'list'>, a=[1, 2, 3] type(a)=<type 'list'>, a=[1, 2, 3] type(a)=<type 'list'>, a=[1, 2, 3] type(a)=<type 'list'>, a=[1, 2, 3] type(a)=<type 'list'>, a=[1, 2, 3] type(a)=<type 'list'>, a=[1, 2, 3] 

and returns:

array([ 6.        ,  5.4       ,  4.90909091,  4.5       ,  4.15384615,         3.85714286,  3.6       ,  3.375     ,  3.17647059,  3.        ]) 

2. currying

since passing same list function call each item in array, can store list directly function currying before applying vectorization:

def curry_f(a):     def f_curried(x):         return f(x, a)  # using definition above     return f_curried  f2 = np.vectorize(curry_f(p)) f2(x) 

prints:

type(a)=<type 'list'>, a=[1, 2, 3] type(a)=<type 'list'>, a=[1, 2, 3] type(a)=<type 'list'>, a=[1, 2, 3] type(a)=<type 'list'>, a=[1, 2, 3] type(a)=<type 'list'>, a=[1, 2, 3] type(a)=<type 'list'>, a=[1, 2, 3] type(a)=<type 'list'>, a=[1, 2, 3] type(a)=<type 'list'>, a=[1, 2, 3] type(a)=<type 'list'>, a=[1, 2, 3] type(a)=<type 'list'>, a=[1, 2, 3] 

and returns:

array([ 6.        ,  5.4       ,  4.90909091,  4.5       ,  4.15384615,         3.85714286,  3.6       ,  3.375     ,  3.17647059,  3.        ]) 

p.s. may wish have @ np.frompyfunc -- similar vectorize(), works @ lower level.