将字符串从 pandas dataframe 转换为列表 - python

Question

I have a pandas dataframe which has a column structured as well:

  sequences
-------------
[(1838, 2038)]
[]
[]
[(809, 1090)]

I'need to loop row by row, so I structured the loop as well:

for index, row in df.iterrows():
    true_anom_seq = json.loads(row['sequences'])

What I wanna do is create a nested loop like [[1838, 2038], [], [], [809, 1090]] so I can iterate through it. The problem is that the code I wrote gives me the error:

JSONDecodeError: Expecting value: line 1 column 2 (char 1)

I also tried to print row['sequences'][0] and it gives me [ , so it is reading it as a string.

How can I convert this string to a list?

Answer 1

Use ast.literal_eval to convert strings to list/dict/...:

from ast import literal_eval

>>> literal_eval('[1,2,3]')
[1,2,3]

Answer 2

import pandas as pd
import re
col = {'index': [1,2,3,4], 'sequence':['[(1838, 2038)]', '[]', '[]', '[(809, 1090)]']}
new_sequence = []
new_df = pd.DataFrame(col)
for index, row in new_df.iterrows():
    one_item = []
    true_anom_seq = re.findall(r'\d+', row['sequence'])
    for match in true_anom_seq:
        one_item.append(match)
    new_sequence.append(one_item)
print(new_sequence)

Answer 3

No need to iterate through the dataframe itself nor use regex. Just apply the literal_eval function to each row in the sequence column and wrap it as a list:

from ast import literal_eval
import pandas as pd

col = {'index': [1,2,3,4], 'sequence':['[(1838, 2038)]', '[]', '[]', '[(809, 1090)]']}
new_sequence = []
new_df = pd.DataFrame(col)

list(new_df.sequence.apply(literal_eval))
[[(1838, 2038)], [], [], [(809, 1090)]]