Break a column of dictionaries to new columns

user_3711985 0
Viewed 56

In my dataframe, there is a column of dictionaries:

ID name value stats
{'mean': 154.0, 'median': 154.0, 'std': 0.0}
{'mean': 131.19, 'median': 93.68, 'std': 53.04}

I need to break down that column as new columns

ID name value mean median std
154.0 154.0 0.0
131.19 93.68 53.04

I tried to use pd.json_normalize as follow:

df2 = pd.json_normalize(df['stats'])
df2

But this way I lose the indexes in df2 and can't join them to add new columns to df. How should I do it?

dataframe python
1 Answers

You can try applying a pd.Series on each dictionary , it would convert as individual dataframe, followed by merging

pd.concat([df, df['col'].apply(pd.Series)], axis=1).drop('col',axis=1)

other approach

df.merge(df['col'].apply(pd.Series), left_index=True, right_index=True, how='outer').drop('col',axis=1)

Example

df = pd.DataFrame()
df['col'] = [{'mean': 154.0, 'median': 154.0, 'std': 0.0},
             {'mean': 131.19, 'median': 93.68, 'std': 53.04}]
df['some'] =1

Prior

    col some
0   {'mean': 154.0, 'median': 154.0, 'std': 0.0}    1
1   {'mean': 131.19, 'median': 93.68, 'std': 53.04} 1

OUt:

    some    mean    median  std
0   1   154.00  154.00  0.00
1   1   131.19  93.68   53.04
user_82080060
Related