Python invalid syntax when using eval(strClass)

Question

I have a list of class names as strings, and if I say fe. print(cNames[0]) it will result into Foo since cNames[0] = "Foo". Now if I wanted to compare a class string with a class, I would do this:

if eval(cNames[0]) in classes:
    foo = eval(cNames[0])()

How ever, it's giving me invalid syntax error:

  File ".\testing.py", line 54, in convert
    print("Class: %s" %eval(cNames[0]))
  File "<string>", line 1
    <class 'testing.Foo'>
    ^

SyntaxError: invalid syntax

And I know I'm only printing the class name, but it didn't work for if sentence either, so I tried to do it with a print. No effect

EDIT: So the problem appeared to be me calling eval("<class 'testing.Foo'>") instead of eval("Foo"). Now this leads into few problems, I'm working with sqlite3 database and I'm trying to save a string as following: <class>:i for example Foo:53, but instead I'm doing <class 'testing.Foo'>:53, and that's why calling eval on it wouldn't work. How could I turn <class 'testing.Foo'> (not a string, the class itself) into string "Foo"?

Why's that? How should I compare class string with a class then? — , Oct 14 '12 at 10:42
also, is this py3? you use print as a function, not a statement — unddoch, Oct 14 '12 at 10:44
If you have a list of class names as strings, `cNames[0] in classes` should work. — Markus Unterwaditzer, Oct 14 '12 at 10:44
It obviously is py3. I tried doing `if getattr(classes, cNames[0]):` and I got this error: `AttributeError: 'list' object has no attribute ''` — , Oct 14 '12 at 10:45
@Mahi so you don't have a list of class names, but a list of their string representations. — Markus Unterwaditzer, Oct 14 '12 at 10:45
@MarkusUnterwaditzer I have a list of classes called `classes` and then I have a list of class names as strings called `cNames[0]` — , Oct 14 '12 at 10:45
@Mahi Print ``cNames`` - it's not what you think it is. It appears to be ``['', ...]``. — Gareth Latty, Oct 14 '12 at 10:46
Can you post the full code please? Also, you can't use gettattr on lists, because the objects they store are not their attributes. You can use L[L.index(obj)] or something as alternative — unddoch, Oct 14 '12 at 10:48

score 3 · Answer 1 · answered Oct 14 '12 at 10:44

3

I'm pretty sure what is happening here is you are doing the eval() twice - and that cNames[0] is the representation you get of the class when you print it. That is, you are doing:

eval("<class 'testing.Foo'>")

I will say, however, this is a bad way of doing it. Store a dictionary which goes from name to class and use that - it's the clearer, more reliable, safer, faster method (you can generate this dictionary if you really feel it's needed).

answered Oct 14 '12 at 10:44

Gareth Latty

86,389
17
178
183

2

You should do the `eval` not even once. – Markus Unterwaditzer Oct 14 '12 at 10:45
Of course - my first comment on the question was saying that, but I'm explaining why it doesn't work - whether it's a terrible idea or not. – Gareth Latty Oct 14 '12 at 10:46

score 2 · Accepted Answer · answered Oct 14 '12 at 11:04

Assuming you have a list containing your classes, you can do this quite tidily without using eval

class Foo(object):
    pass

class Thing(object):
    pass

# List of interesting classes
classes = [Foo, Thing]

# Construct a dictionary where the key is the Classes __name__ attribute
classByNames = {cls.__name__:cls for cls in classes}

# Example usage
cName = 'Foo'
if cName in classByNames:
    inst = classByNames[cName]()
    print(inst) # <__main__.Foo object at ...>
else:
    print("Unknown class %s" % cName)

(only tested the above in Python v2.7, but should work fine in v3 also)

Python invalid syntax when using eval(strClass)

2 Answers2