pandas: keyword error while accessing column

Question

I have a pandas dataframe which at beginning looks like this

Now I did some computation with the dataframe and the code is as follows

df_sorted = (df.groupby(df['location_id'])['users'].nunique()).to_frame()
df_sorted = df_sorted.sort_values(['users'], ascending=False)

now the dataframe looks like this

but now I cannot access location_id from new dataframe with following code

df_sorted['location_id']

I get error saying KeyError: 'location_id' However when i do

df_sorted['users]

I get entire dataframe like

Any idea why??

You should stop at groupby() and see what your dataframe was after that step :D — Phung Duy Phong
– Phung Duy Phong, Commented May 8, 2017 at 11:50
As @zipa says, location_id is now an index, so you either access the index by df_sorted.index or you make a new column df.sorted['location_id'] = df_sorted.index — VinceP
– VinceP, Commented May 8, 2017 at 11:51
When in doubt inspect df.columns, or better yet df.columns.tolist() to spot blank spaces. — IanS
– IanS, Commented May 8, 2017 at 11:52

zipa · Accepted Answer · 2017-05-08 11:51:44Z

2

location_id becomes index after groupby(). What you can do is use:

df_sorted = df_sorted.sort_values(['users'], ascending=False).reset_index()

This will return to previous indexing and turn location_id into column.

answered May 8, 2017 at 11:51

zipa

28k6 gold badges45 silver badges62 bronze badges

Sign up to request clarification or add additional context in comments.

1 Answer 1