Pandas — MultiIndex

Create MultiIndex

import pandas as pd

arrays = [['A', 'A', 'B', 'B'], ['one', 'two', 'one', 'two']]
index = pd.MultiIndex.from_arrays(arrays, names=['first', 'second'])
df = pd.DataFrame({'data': [1, 2, 3, 4]}, index=index)

Select with MultiIndex

# Select level
print(df.loc['A'])

# Select specific
print(df.loc[('A', 'one')])

Swap levels

df = df.swaplevel()

Sort index

df = df.sort_index()

Stack/Unstack

# Unstack level
unstacked = df.unstack(level='second')

# Stack
stacked = unstacked.stack()

Mini Practice

1. Create MultiIndex
2. Select by level
3. Swap levels
4. Use stack/unstack

Up Next

Continue with Resample - Time resampling.