Most probably a stupid question, but I still want to ask

Rustom Mody rustompmody at gmail.com
Mon Apr 11 01:18:08 EDT 2016


On Monday, April 11, 2016 at 10:17:13 AM UTC+5:30, Stephen Hansen wrote:
> On Sun, Apr 10, 2016, at 09:03 PM, Fillmore wrote:
> > and the (almost always to be avoided) use of eval()
> 
> FWIW, there's ast.literal_eval which is safe and there's no reason to
> avoid it.

Its error reporting is clunky:

>>> from ast import literal_eval as le
>>> le("(1)")
1
>>> le("(1,)")
(1,)
>>> le("1")
1
>>> le("{1:x}")
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/lib/python2.7/ast.py", line 80, in literal_eval
    return _convert(node_or_string)
  File "/usr/lib/python2.7/ast.py", line 63, in _convert
    in zip(node.keys, node.values))
  File "/usr/lib/python2.7/ast.py", line 62, in <genexpr>
    return dict((_convert(k), _convert(v)) for k, v
  File "/usr/lib/python2.7/ast.py", line 79, in _convert
    raise ValueError('malformed string')
ValueError: malformed string


> You'll still have to deal with the fact that a single string
> on a line will return a string while multiples will return a tuple, but
> how you're doing that (checking the type of the result) is fine.

Yes...
0 and 1 length tuples are special cases even if we argue till we are blue in 
the face

To Fillmore:

Go via lists and the special cases can be obviated:
>>> def br(s): return "[" + s + "]"
>>> s0=""
>>> s1="1"
>>> s2="1,2"
>>> le(br(s0))
[]
>>> le(br(s1))
[1]
>>> le(br(s2))
[1,2]

Then tuplify (if you so wish)
>>> tuple(le(br(s0)))
()
>>> tuple(le(br(s1)))
(1,)
>>> tuple(le(br(s2)))
(1, 2)



More information about the Python-list mailing list