A high-performance, memory-efficient library for numeric data manipulation with native missing data support.

• Native support for missing data
• Fast performance (at least as good as Pandas)
• Memory-efficient storage optimized for large datasets
• Sparse array support for reduced memory footprint
• Serialization & I/O support

To install numdrift, use pip:

pip install numdrift

import numdrift as nd

# Create a Numdrift array
arr = nd.mdarray([1, 2, 3, None, 5])

# Perform operations on the array
result = arr.mean()
print(result)