awkward 0.12.3

Manipulate arrays of complex data structures as easily as Numpy.

Calculations with rectangular, numerical data are simpler and faster in Numpy than traditional for loops. Consider, for instance,

all_r = []
for x, y in zip(all_x, all_y):
    all_r.append(sqrt(x**2 + y**2))

versus

all_r = sqrt(all_x**2 + all_y**2)

Not only is the latter easier to read, it’s hundreds of times faster than the for loop (and provides opportunities for hidden vectorization and parallelization). However, the Numpy abstraction stops at rectangular arrays of numbers or character strings. While it’s possible to put arbitrary Python data in a Numpy array, Numpy’s dtype=object is essentially a fixed-length list: data are not contiguous in memory and operations are not vectorized.

Awkward-array is a pure Python+Numpy library for manipulating complex data structures as you would Numpy arrays. Even if your data structures

• contain variable-length lists (jagged/ragged),
• are deeply nested (record structure),
• have different data types in the same list (heterogeneous),
• are masked, bit-masked, or index-mapped (nullable),
• contain cross-references or even cyclic references,
• need to be Python class instances on demand,
• are not defined at every point (sparse),
• are not contiguous in memory,
• should not be loaded into memory all at once (lazy),

this library can access them as columnar data structures, with the efficiency of Numpy arrays. They may be converted from JSON or Python data, loaded from “awkd” files, HDF5, Parquet, or ROOT files, or they may be views into memory buffers like Arrow.

Installation

Install awkward like any other Python package:

pip install awkward                       # maybe with sudo or --user, or in virtualenv
pip install awkward-numba                 # optional: integration with and optimization by Numba

or install with conda:

conda config --add channels conda-forge   # if you haven't added conda-forge already
conda install awkward
conda install awkward-numba               # optional: integration with and optimization by Numba

The base awkward package requires only Numpy (1.13.1+), but awkward-numba additionally requires Numba.

Recommended packages:

• pyarrow to view Arrow and Parquet data as awkward-arrays
• h5py to read and write awkward-arrays in HDF5 files
• Pandas as an alternative view