使用熊猫取消嵌套数组列

Question

My co-worker has handed me a dataset containing a column with nested data:

id | ... | x
 0 | ... | [['a',3.0],['b',5.4]]
 1 | ... | [['a',1.3],['b',7.6]]
 2 | ... | [['b',2.4],['a',8.8]]
 : | ... | :                   :

I am trying to parse the column to multiple columns to get something like this:

id | ... | a   | b
 0 | ... | 3.0 | 5.4
 1 | ... | 1.3 | 7.6
 2 | ... | 8.8 | 2.4
 : | ... | :   | :

unfortunately I can't seem to find the way to do this using pandas. I know that nesting data this way is to be discouraged but unfortunately my co-worker doesn't perceive this to be a problem so this is how I'm spending my weekend.

Does anyone have experience handling this type of problem?

Answer 1

I think this should work:

records = [ dict(row) for row in df['x'].values ]
new_df = pd.DataFrame.from_records(records, columns = records[0].keys(), index = df.index)

(I did not test it, so it might required some changes)