Apply exec function to pandas DataFrame

Question

I have following pandas DataFrame to which I need to apply the exec function to all the rows.

import pandas as pd
var = 2.5
df = pd.DataFrame(["var*1", "var*3", "var*5"])

Expected result

The expected result I need is:

      0
0   2.5
1   7.5
2  12.5

Using exec function with apply

However if I do the following:

df.apply(exec)

or even:

df.apply(lambda x:exec(x))

I get the following error:

TypeError: exec() arg 1 must be a string, bytes or code object

Using map with lambda function

Other option I tried was:

>>> df[0].map(lambda x:exec(x))
0    None
1    None
2    None
Name: 0, dtype: object

But it's not the result I expect.

Defining a function

I even tried to define a function and then apply it:

def execute(x):
    exec("a = " + str(x))
    return a

df[0].apply(execute)

But result is following error:

NameError: name 'var' is not defined

Question

How can I then get the expected result?

Answer 1

Take a look at pandas.eval. You can reference local variables using @var syntax.

As for arbitrary code execution, you should explicitly pass local variables as dict and use eval to get a return value, e.g. eval("a + 2", {"a": 1})

As for your question:

df[0].apply(eval, args=({}, {"var": var}))