Python Pandas - Set levels in a MultiIndex

A MultiIndex in Pandas is a hierarchical index that allows multiple levels of indexing. The set_levels() method replaces the current level values with new ones while preserving the structure.

Creating a MultiIndex

First, let's create a MultiIndex from arrays and examine its structure ?

import pandas as pd

# Create arrays for MultiIndex levels
arrays = [[1, 2, 3, 4], ['John', 'Tim', 'Jacob', 'Chris']]

# Create MultiIndex with named levels
multiIndex = pd.MultiIndex.from_arrays(arrays, names=('ranks', 'student'))

print("Original MultiIndex:")
print(multiIndex)
print("\nCurrent levels:")
print(multiIndex.levels)

Original MultiIndex:
MultiIndex([(1,  'John'),
            (2,   'Tim'),
            (3, 'Jacob'),
            (4, 'Chris')],
           names=['ranks', 'student'])

Current levels:
[[1, 2, 3, 4], ['Chris', 'Jacob', 'John', 'Tim']]

Setting New Levels

Use set_levels() to replace existing level values with new ones ?

import pandas as pd

arrays = [[1, 2, 3, 4], ['John', 'Tim', 'Jacob', 'Chris']]
multiIndex = pd.MultiIndex.from_arrays(arrays, names=('ranks', 'student'))

# Set new levels - replaces values but keeps structure
new_multiIndex = multiIndex.set_levels([['A', 'B', 'C', 'D'], ['Alice', 'Bob', 'Charlie', 'David']])

print("MultiIndex with new levels:")
print(new_multiIndex)
print("\nNew level values:")
print(new_multiIndex.levels)

MultiIndex with new levels:
MultiIndex([('A', 'Alice'),
            ('B',   'Bob'),
            ('C', 'Charlie'),
            ('D',  'David')],
           names=['ranks', 'student'])

New level values:
[['A', 'B', 'C', 'D'], ['Alice', 'Bob', 'Charlie', 'David']]

Setting Levels for Specific Level Only

You can also set levels for a specific level using the level parameter ?

import pandas as pd

arrays = [[1, 2, 3, 4], ['John', 'Tim', 'Jacob', 'Chris']]
multiIndex = pd.MultiIndex.from_arrays(arrays, names=('ranks', 'student'))

# Set levels only for the first level (ranks)
updated_multiIndex = multiIndex.set_levels(['First', 'Second', 'Third', 'Fourth'], level=0)

print("Updated MultiIndex (first level only):")
print(updated_multiIndex)

Updated MultiIndex (first level only):
MultiIndex([('First', 'John'),
            ('Second',  'Tim'),
            ('Third', 'Jacob'),
            ('Fourth', 'Chris')],
           names=['ranks', 'student'])

Key Points

• set_levels() returns a new MultiIndex object; it doesn't modify the original
• New level values must have the same length as existing levels
• The structure and codes remain unchanged, only level values are replaced
• Use the level parameter to update specific levels only

Conclusion

The set_levels() method allows you to replace MultiIndex level values while preserving the hierarchical structure. Use it when you need to rename or update level values without changing the index organization.