我有一个数据框,其中在一列中,每一行的数据都是这样的字符串:
[[25570], [26000]]
我希望系列中的每个条目成为整数列表。
IE:
[25570,26000] ^^ int int
到目前为止,我可以将其保存到字符串列表中,但保留空白:
s = s.str.replace("[","").str.replace("]","")
s = s.str.replace(" ","").str.split(",")
数据框字典:
f = {'chunk': {0: '[72]',
1: '[72, 68]',
2: '[72, 68, 65]',
3: '[72, 68, 65, 70]',
4: '[72, 68, 65, 70, 67]',
5: '[72, 68, 65, 70, 67, 74]',
6: '[68]',
7: '[68, 65]',
8: '[68, 65, 70]',
9: '[68, 65, 70, 67]'},
'chunk_completed': {0: '[25570]',
1: '[26000]',
2: '[26240]',
3: '[26530]',
4: '[26880]',
5: '[27150]',
6: '[26000]',
7: '[26240]',
8: '[26530]',
9: '[26880]'},
'chunk_id': {0: '72',
1: '72-68',
2: '72-68-65',
3: '72-68-65-70',
4: '72-68-65-70-67',
5: '72-68-65-70-67-74',
6: '68',
7: '68-65',
8: '68-65-70',
9: '68-65-70-67'},
'diffs_avg': {0: nan,
1: 430.0,
2: 335.0,
3: 320.0,
4: 327.5,
5: 316.0,
6: nan,
7: 240.0,
8: 265.0,
9: 293.3333333333333},
'sd': {0: nan,
1: nan,
2: 134.35028842544406,
3: 98.48857801796105,
4: 81.80260794538685,
5: 75.3657747256671,
6: nan,
7: nan,
8: 35.355339059327385,
9: 55.075705472861024},
'timecodes': {0: '[[25570]]',
1: '[[25570], [26000]]',
2: '[[25570], [26000], [26240]]',
3: '[[25570], [26000], [26240], [26530]]',
4: '[[25570], [26000], [26240], [26530], [26880]]',
5: '[[25570], [26000], [26240], [26530], [26880], [27150]]',
6: '[[26000]]',
7: '[[26000], [26240]]',
8: '[[26000], [26240], [26530]]',
9: '[[26000], [26240], [26530], [26880]]'}}
答案 0 :(得分:2)
尝试
f = pd.DataFrame().from_dict(s, orient='index')
f.columns = ['timecodes']
f['timecodes'].apply(lambda x: [a[0] for a in eval(x) if a])
输出
Out[622]:
0 [25570]
1 [25570, 26000]
2 [25570, 26000, 26240]
3 [25570, 26000, 26240, 26530]
4 [25570, 26000, 26240, 26530, 26880]
5 [25570, 26000, 26240, 26530, 26880, 27150]
6 [26000]
7 [26000, 26240]
8 [26000, 26240, 26530]
9 [26000, 26240, 26530, 26880]
10 [26000, 26240, 26530, 26880, 27150]
11 [26240]
12 [26240, 26530]
13 [26240, 26530, 26880]
14 [26240, 26530, 26880, 27150]
15 [26530]
16 [26530, 26880]
17 [26530, 26880, 27150]
18 [26880]
19 [26880, 27150]
Name: 0, dtype: object