Overview

Vaex 4.17, released on July 21, 2023, is the last active release of the out-of-core DataFrame library.

Main Features

Last active release

Vaex 4.17 brings final fixes and improvements. Vaex processes multi-GB DataFrames without loading them entirely into memory using memory-mapping and lazy evaluation.

python
import vaex

# Open a multi-GB HDF5 file
# (memory-mapped, not loaded into memory)
# df = vaex.open('big_data.hdf5')

# Example with in-memory data
df = vaex.from_dict({
    'x': [1, 2, 3, 4, 5],
    'y': [10, 20, 30, 40, 50],
})

# Lazy evaluation
df['z'] = df.x + df.y
print(df.mean(df.z))  # 33.0

Sources