Python: why does list comprehension produce a generator?

Question

a=[['1','3','2'],['11','22','33']]
k = [(float(a[i][j]) for j in range(0,3)) for i in range(0,2)]
>>> print k
[<generator object <genexpr> at 0x7f1a9d568f50>, <generator object <genexpr> at 0x7f1a9d568fa0>]

but I want to get [(1,3,2),(11,22,33)] why does list comprehension produce a generator?

Answer 1

You have a generator expression ((x for x in ...)) inside a list comprehension ([x for x in ...]). This will return a list of generator objects. Change the code like so

a = [['1','3','2'],['11','22','33']]
k = [[float(a[i][j]) for j in range(0,3)] for i in range(0,2)]

print(k)
# [[1.0, 3.0, 2.0], [11.0, 22.0, 33.0]]

Answer 2

You are using a generator expression in your list comprehension expression:

(float(a[i][j]) for j in range(0,3))

If you wanted that to execute like a list comprehension, make that a list comprehension too:

[[float(a[i][j]) for j in range(3)] for i in range(2)]

If you need those to be tuples, then you'll have to explicitly call tuple():

[tuple(float(a[i][j]) for j in range(3)) for i in range(2)]

The tuple() callable will drive the generator expression to produce a tuple of those values. There is no such thing as a tuple comprehension, otherwise.

Rather than use ranges, you can loop over a and the nested lists directly:

[tuple(float(v) for v in nested) for nested in a]

Demo:

>>> a=[['1','3','2'],['11','22','33']]
>>> [tuple(float(v) for v in nested) for nested in a]
[(1.0, 3.0, 2.0), (11.0, 22.0, 33.0)]