CODEWITHSUNDEEP
python
Stefan Falk
Stefan Falk

Reputation: 25397

TypeError: cannot perform reduce with flexible type (list of lists of string)

I have a list r and a list l

r = get_r()
l = [['a', 'b'], ['a'], ['c']]

print('-')
[print(type(el)) for el in l];
print('-')
[print(type(el)) for el in r];

will give us

-
<class 'list'>
<class 'list'>
<class 'list'>
-
<class 'list'>
<class 'list'>
<class 'list'>

So let's print r and l

print(l)
print(r)

will give us

[['a', 'b'], ['a'], ['c']]
[['Schadenersatzrecht'], ['Abgabenrecht, Finanzrecht und Steuerrecht; Verfahrensrecht'], ['Europarecht']]

But now as I np.sum those lists up:

np.sum(l)
np.sum(r)

I get:

['a', 'b', 'a', 'c']

but also:

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-158-0d0bce575128> in <module>()
----> 1 np.sum(r)

~/miniconda3/lib/python3.6/site-packages/numpy/core/fromnumeric.py in sum(a, axis, dtype, out, keepdims)
   1832             return sum(axis=axis, dtype=dtype, out=out, **kwargs)
   1833     return _methods._sum(a, axis=axis, dtype=dtype,
-> 1834                          out=out, **kwargs)
   1835 
   1836 

~/miniconda3/lib/python3.6/site-packages/numpy/core/_methods.py in _sum(a, axis, dtype, out, keepdims)
     30 
     31 def _sum(a, axis=None, dtype=None, out=None, keepdims=False):
---> 32     return umr_sum(a, axis, dtype, out, keepdims)
     33 
     34 def _prod(a, axis=None, dtype=None, out=None, keepdims=False):

TypeError: cannot perform reduce with flexible type

I totally not understand what the problem is. Even printing the type in the lists of the elements in those lists [print(type(el[0])) for el in r]; will give me <class 'str'> as a result. The types are identical - at least from what I can tell yet np.sum won't work on r.

I am not sure if one can tell what the issue is from what you can see here. The only information I can provide in addition is that r comes from a parsed JSON - however, I don't see why this would affect the outcome.

Upvotes: 2

Views: 1427

Answers (1)

Moses Koledoye
Moses Koledoye

Reputation: 78556

Using np.sum for flattening your lists is a bad idea. Before np.sum is performed, both lists are coerced to np array types, which have different dtypes (although that is slightly aside the reason for failure).

np.array(l).dtype != np.array(r).dtype

You generally don't want to use np.sum for this.

Simply flatten your list of lists the Pythonic way using list comprehensions:

l = [x for lst in l for x in lst]
r = [x for lst in r for x in lst]

Upvotes: 2

Related Questions