How to convert DataFrame?-CodePudding

I have a DataFrame.

id   actions
1   [{"action_type":"link_click","value":"1","1d_click":"1","7d_click":"1"},{"action_type":"page_engagement","value":"2","1d_click":"2","7d_click":"2"}]
2   [{"action_type":"link_click","value":"2","1d_click":"2","7d_click":"2"},{"action_type":"page_engagement","value":"2","1d_click":"2","7d_click":"2"}]

I have to convert like this.

id    action_type       value
1     link_click        1
1     page_engagement   2
2     link_click        2
2     page_engagement   2

Is this possible with pandas? If so, how can I manage to do it?

CodePudding user response：

pd.DataFrame((df := df.apply(lambda s: list(map(lambda t: [t['action_type'], t["value"]], s[1])), axis=1).explode()).to_list() , index= df.index)

if your actions column is str:

import ast
pd.DataFrame((df := df.apply(lambda s: list(map(lambda t: [t['action_type'], t["value"]], ast.literal_eval(s[1]))), axis=1).explode()).to_list() , index= df.index)