Python Pandas - Get location for a sequence of labels in a MultiIndex

To get location for a sequence of labels in a MultiIndex, use the MultiIndex.get_locs() method in Pandas. This method returns an array of integer positions where the specified label appears in the MultiIndex.

Creating a MultiIndex

First, let's create a MultiIndex using from_arrays() ?

import pandas as pd

# Create MultiIndex from two arrays
multiIndex = pd.MultiIndex.from_arrays([list('pqrrst'), list('kytssp')])

# Display the MultiIndex
print("The MultiIndex...")
print(multiIndex)

The MultiIndex...
MultiIndex([('p', 'k'),
            ('q', 'y'),
            ('r', 't'),
            ('r', 's'),
            ('s', 's'),
            ('t', 'p')],
           )

Getting Locations with get_locs()

The get_locs() method returns the positions where a label appears. Let's find all locations of label 's' ?

import pandas as pd

# Create MultiIndex
multiIndex = pd.MultiIndex.from_arrays([list('pqrrst'), list('kytssp')])

# Get locations for label 's' in the first level
print("Locations of 's' in MultiIndex:")
locations = multiIndex.get_locs('s')
print(locations)

# Show the levels for reference
print("\nThe levels in MultiIndex:")
print(multiIndex.levels)

Locations of 's' in MultiIndex:
[4]

The levels in MultiIndex:
[['p', 'q', 'r', 's', 't'], ['k', 'p', 's', 't', 'y']]

Finding Multiple Occurrences

Let's create an example where a label appears multiple times to better understand get_locs() ?

import pandas as pd

# Create MultiIndex with repeated labels
multiIndex = pd.MultiIndex.from_arrays([list('abacad'), list('xyzxyz')])

print("MultiIndex with repeated labels:")
print(multiIndex)

# Get locations for label 'a' (appears multiple times)
print("\nLocations of 'a':")
locations = multiIndex.get_locs('a')
print(locations)

# Verify by showing actual values at those positions
print("\nValues at those locations:")
for loc in locations:
    print(f"Position {loc}: {multiIndex[loc]}")

MultiIndex with repeated labels:
MultiIndex([('a', 'x'),
            ('b', 'y'),
            ('a', 'z'),
            ('c', 'x'),
            ('a', 'y'),
            ('d', 'z')],
           )

Locations of 'a':
[0 2 4]

Values at those locations:
Position 0: ('a', 'x')
Position 2: ('a', 'z')
Position 4: ('a', 'y')

Key Points

• The get_locs() method searches the first level of the MultiIndex by default
• It returns a NumPy array of integer positions
• If a label appears multiple times, all positions are returned
• If a label doesn't exist, it raises a KeyError

Conclusion

Use MultiIndex.get_locs() to find all positions where a specific label appears in a MultiIndex. This method is useful for locating and accessing multiple rows that share the same index label.