如何在pandas数据帧中反转.astype（str）？

Question

I had to remove duplicate rows in my dataframe which had list values in it.

So I used

pd_data['douban_info_string'] = pd_data['douban_info'].astype(str)

Where 'douban_info_string' had list values.

But now I need this list to compare with list of another data frame. But the list is changed into string now and I get this error

TypeError: unhashable type: 'list'

Answer 1

Use pandas.eval :

df = pd.DataFrame({'info':[[1,2,3], [4,5,6]]})

df['info_str']=df['info'].astype(str)
df['info_str'][0]
# '[1, 2, 3]'

df['info_str'].apply(pd.eval)[0]
# [1,2,3]

Answer 2

Use apply with an if statement:

df = pd.DataFrame({'info':[[1,2,3], [4,5,6], 'str224']})
df['info_str'] = df['info'].astype(str)
print(df['info_str'][0])
print(type(df['info_str'][0]))
print(df['info_str'].apply(lambda x: x if x in df['info'].tolist() else pd.eval(x))[0])
print(type(df['info_str'].apply(lambda x: x if x in df['info'].tolist() else pd.eval(x))[0]))

Output:

[1, 2, 3]
<class 'str'>
[1 2 3]
<class 'numpy.ndarray'>

Answer 3

Try this

pd_data['douban_info_string_list'] = pd_data['douban_info_string'].map(lambda x: x.replace('[', '').replace(']', '').split(','))

Hope it helps.