Pandas: After using transpose old index name does not vanish when using dataFrame.index.set_names

user_14824451 0
Viewed 65

When transposing a pandas Dataframe which has a named index, the previous index name ( for the index in the first column ) appears as the first entry for the column names.

Example: original DataFrame df_1

1800 1801 1802
country
Germany 38.4 38.4 38.4 
df_2 = df_1.T

df_2 is

country Germany
1800 38.4
1801 38.4
1802 38.4

First question. Why is country now the index for 1800, 1801 etc. and is there a better tranpose option to avoid this?

When trying to rename the index with

df_2.index.set_names(["year"],inplace=True)

the following result is shown

country Germany
year
1800 38.4
1801 38.4
1802 38.4

Question two. Why is country still there and how to remove it?

pandas transform dataframe
2 Answers

Country is the name of df_1 index, you can check:

df_check = df_1.rename_axis(columns=['year'], index=['country'])
print(df_check.index.name, df_check.columns.name)

So you could do:

df_1 = df_1.rename_axis(index=None)
user_118842370

First question. Why is country now the index for 1800, 1801 etc. and is there a better tranpose option to avoid this?

'country' is the name/ label of the column axis since it was the name of the index of the DataFrame before transposing.

df_2.index.set_names(["year"],inplace=True)

Question two. Why is country still there and how to remove it?

Because you are only changing the index name of the transposed DataFrame. The name of the column axis ('country') is left unchanged.

How to remove it:

You can use DataFrame.rename_axis to change the name of both axes (index and column) at the same time. If you want to remove the name of a given axis, just pass None.

For instance,

# or df_2 = df1.T.rename_axis(index='year', columns=None) if you prefer
>>> df_2 = df_1.rename_axis(index=None, columns='year').T
>>> df_2

      Germany
year         
1800     38.4
1801     38.4
1802     38.4
user_171206920
Related