pandas pivot_table keep index

set_index with append

df.set_index(['day_bucket', 'label'], append=True) \
  .rename_axis([None, None, None]).squeeze().unstack()


              birds
0 2011-01-21      4
1 2011-01-22      0
2 2011-01-23      7
3 2011-01-24      3

In the meantime, I also came up with a result

pd.pivot_table(df, values='numeric_value', 
                   index=[df.index.values,'day_bucket'],
                   ,columns='label').reset_index('day_bucket')

label  day_bucket  mortality_birds
0      2011-01-21                4
1      2011-01-22                0
2      2011-01-23                7
3      2011-01-24                3

just for supplement:

pt2 = pt.rename_axis(None, axis=1).reset_index()
print(pt2)

   day_bucket  birds
0  2011-01-21      4
1  2011-01-22      0
2  2011-01-23      7
3  2011-01-24      3

---

the dataframe used for it is:

df = pd.DataFrame({'day_bucket': ['2011-01-21', '2011-01-22', '2011-01-23', '2011-01-24'],
                   'label': ['birds', 'birds', 'birds', 'birds'],
                   'num_value': [4, 0, 7, 3]})

pt = pd.pivot_table(df,
                    values='num_value',
                    index='day_bucket',
                    columns='label',
                    aggfunc=np.sum)
print(pt, '\n')

label       birds
day_bucket       
2011-01-21      4
2011-01-22      0
2011-01-23      7
2011-01-24      3