Here's my problem. I have a dataframe with x columns and y lines. Some columns are actually lists. I want to transform those columns to multiple columns containing single values.
An example speaks by itself :
My dataframe :
ans_length ans_unigram_numbers ... levenshtein_dist que_entropy
0 [19, 14] [12, 8] ... 9.00 3.189898
1 [19] [12] ... 4.00 3.189898
2 [0] [0] ... 170.00 4.299996
3 [0] [0] ... 170.00 4.303341
4 [0] [0] ... 170.00 4.304335
5 [0] [0] ... 170.00 4.311820
28 [56] [23] ... 24.00 4.110291
29 [0] [0] ... 56.00 4.181720
... ... ... ... ... ...
1976 [24] [11] ... 24.00 3.084963
1977 [24] [11] ... 24.00 3.084963
1992 [31, 24, 32, 28] [14, 15, 17, 11] ... 18.75 3.292770
1993 [31, 24, 32, 28] [14, 15, 17, 11] ... 18.75 3.292770
[1998 rows x 9 columns]
What I expect :
ans_length_0 ans_length_1 ans_length_2 ans_length_3 \
0 19 14
1 19
2 0
3 0
4 0
5 0
28 56
29 0
1976 24
1977 24
1992 31 24 32 28
1993 31 24 32 28
ans_unigram_numbers_0 ans_unigram_numbers_1 ans_unigram_numbers_2 ans_unigram_numbers_3 \
12 8
12
0
0
0
0
23
0
11
11
14 15 17 11
14 15 17 11
levenshtein_dist que_entropy
9 3.189898
4 3.189898
170 4.299996
170 4.303341
170 4.304335
170 4.31182
24 4.110291
56 4.18172
24 3.084963
24 3.084963
18.75 3.29277
18.75 3.29277
The newly generated columns should take the name of the old one, adding an index at the end of it.
I think you can use:
cols = ['ans_length','ans_unigram_numbers']
df1 = pd.concat([pd.DataFrame(df[x].values.tolist()).add_prefix(x) for x in cols], axis=1)
df = pd.concat([df1, df.drop(cols, axis=1)], axis=1)
Based on @jezrael answer, I created a function that do what is asked, from a given dataframe and a given list of columns :
def flattencolumns(df1, cols):
df = pd.concat([pd.DataFrame(df1[x].values.tolist()).add_prefix(x) for x in cols], axis=1)
return pd.concat([df, df1.drop(cols, axis=1)], axis=1)
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.