'DataFrameGroupBy' object has no attribute 'set_index'

Question

I have a dataFrame data. After grouping and resetting the index, I am not able to set my Date column as Index.

data = data.groupby('Payment Date ')
data['Payment Amount '].sum().reset_index()

This is my output resetting index

 data = data.set_index('Payment Date ', inplace  = True)
 data.index

Error:

AttributeError                            Traceback (most recent call last)
<ipython-input-12-581b0b0bf83f> in <module>
----> 1 data = data.set_index('Payment Date ', inplace  = True)
      2 data.index

c:usersnitroappdatalocalprogramspythonpython37-32libsite-packagespandascoregroupbygroupby.py in __getattr__(self, attr)
    702 
    703         raise AttributeError(
--> 704             f"'{type(self).__name__}' object has no attribute '{attr}'"
    705         )
    706 

AttributeError: 'DataFrameGroupBy' object has no attribute 'set_index'

Asked By: Karanpreet Singh

||

Source

Answer 1

I think I understood what you wanted to do and what you did not understand (mainly about the way to modifiy objects with pandas). I assume that you wanted to:

compute your aggregation by payment date in data
and then set its index to ‘Payment date’ field

Short answer: if you want to have this result into data, simply execute:

data = data.groupby('Payment date ')['Payment amount'].sum().to_frame()

‘Payment date ‘ will be your new index, to_frame prevents your single column resulting dataframe to be squeezed into a pandas Series (which I think was your first intention to avoid, resetting your index to then set it back).

Let’s dive into your code.

First line

data = data.groupby('Payment Date ')

First line is ok, but might not do exactly what you want. You are taking data, which I assume is a pandas DataFrame and reaffect it a pandas DataFrameGroupBy object. This kind of object does not hold any data, you can see it simply as a mapping between index(s) of your original DataFrame and associated groups (here, payment dates).

Anyway, you got your groupby object into data.

Second line

data['Payment Amount '].sum().reset_index()

This line does nothing. It shows the result of the computation in your Jupyter notebook, but nothing has been changed in data. data is still the same DataFrameGroupBy object.

Third line

data = data.set_index('Payment Date ', inplace  = True)

An exception is raised, saying that a DataFrameGroupBy objet has no set_index method. This is because data has not been changed by your second line of code.
Even so, I would encourage you to avoid using inplace=True anytime in your code. You should always go with explicit reassignements.

Your code could look like (if you don’t like the short answer above):

data = data.groupby('Payment date ')
data = data['Payment amount'].sum().reset_index()
data = data.set_index('Payment date ')  # No inplace=True!

Answered By: Pierre Massé

'DataFrameGroupBy' object has no attribute 'set_index'

Question:

Answers: