How to explode a struct index in pandas series

Question

I have a pandas Series where the index is a StructType. I want to explode this index into the structfields.

index	value
(2005-07-03, beta0)	0.997160
(2005-07-03, beta1)	0.037629
(2005-07-03, beta2)	0.037629
(2005-07-03, tau)	2.000000
(2011-11-13, beta0)	0.997160

Here the index is of the structype key: struct<checkin_week:date,parameter:string>

I need to explode it two columns such that I have a column for checkin_week and another column for parameter.

@khaledkoubaa : I changed it to a text format. Not sure why is it that important to answer the question. Could you please explain. — Roopanjali Jasrotia, Sep 23 '22 at 12:33
you need to split this ```(2005-07-03, beta0)``` to ```2005-07-03``` and ```2005-07-03```? I mean split in 2 columns — khaled koubaa, Sep 23 '22 at 12:47

score 0 · Accepted Answer · answered Oct 05 '22 at 14:32

The following will give you a DataFrame with a MultiIndex, in which checkin_week and parameter will be separate levels.

import pandas as pd

df = pd.DataFrame(
    data=(
        0.997160,
        0.037629,
        0.037629,
        2.000000,
        0.997160,
    ),
    index=(
        ("2005-07-03", "beta0"),
        ("2005-07-03", "beta1"),
        ("2005-07-03", "beta2"),
        ("2005-07-03", "tau"),
        ("2011-11-13", "beta0"),
    ),
)

df.index = pd.MultiIndex.from_tuples(df.index)

If you want them to be separate columns (instead of indexes), you should first reset your DataFrame's index:

df = df.reset_index(drop=False)

And then follow the answers of this question

How to explode a struct index in pandas series

1 Answers1