JSON normalize pandas column without losing index

Question

I've got a dataframe as in the image:

I'd like to transform that into a dataframe similar to the following:

But, in the second image, instead of having (0,1,2...) as index, I want to keep card_id as index - or as a column, whatever.

The second image is obtained applying pd.json_normalize in the first df's card_fields column.

I did find something similar in this question, but in there we start with a dictionary, instead of a dataframe, as in my case, so the proposed solution is not directly translatable to my case without doing a lot of manipulation.

`pd.concat([df.reset_index()[['card_id']], pd.json_normalize(df['card_fields'])], axis=1])` — Emma, Feb 17 '22 at 03:15

score 3 · Answer 1 · answered Feb 17 '22 at 02:17

3

try pd.json_normalize(df['card_fields']).set_index(df.index)? df is the dataframe from image 1.

answered Feb 17 '22 at 02:17

CJR

76
4

That was my first try, even before asking the question. Unfortunally, seems like ```pd.json_normalize``` doesn't preserve order, so that wouldn't work. – Davi Alefe Feb 17 '22 at 02:33
1

@DaviAlefe i think `json_normalize` is preserving the order. are you sure it doesn't? – cookiemonster Jul 05 '22 at 19:30
yes, it definitely preserves order – BatyaGG Jun 20 '23 at 08:01

JSON normalize pandas column without losing index

1 Answers1