logo
  • Quickstart
  • What is an Awkward Array?
  • Converting arrays
    • NumPy
    • Python objects
    • JSON
    • ROOT via Uproot
    • Arrow and Parquet
    • Pandas
    • Generic buffers
  • Creating new arrays
    • Arrays of lists
    • Arrays of records
    • Arrays of missing data
    • Arrays of strings
    • Lazy arrays [todo]
    • Unflattening and grouping [todo]
    • ArrayBuilder (easy & general)
    • LayoutBuilder (more control)
    • Direct constructors (fastest)
  • Examining arrays
    • Data type [todo]
    • Single item detail [todo]
    • Listing fields/keys/columns [todo]
    • Simple slicing [todo]
    • Checking validity [todo]
  • Numerical math
    • NumPy functions [todo]
    • Awkward broadcasting [todo]
    • Reducing (sum/min/any/all) [todo]
    • Statistics (mean/var/std) [todo]
    • Using argmin/argmax [todo]
    • On GPUs [todo]
  • Filtering data
    • By number of items [todo]
    • Cuts vs. masks [todo]
    • Slicing lists within arrays [todo]
    • Slices with missing values [todo]
  • Restructuring data
    • Zip/unzip and project [todo]
    • Adding fields to records [todo]
    • Renaming records [todo]
    • Flattening for plots
    • Padding/clipping for machine learning
    • Concatenating and interleaving [todo]
    • Sorting [todo]
  • Combinatorics
    • Cartesian and "n choose k" [todo]
    • Best match between collections [todo]
  • Using arrays in Numba
    • Supported features [todo]
    • Building array output [todo]
  • Specialized behavior
    • Subclassing Array/Record [todo]
    • Overriding NumPy functions [todo]
    • In Numba [todo]
    • For physics: Lorentz vectors [todo]
  • Differentiation using JAX

Numerical mathΒΆ

  • NumPy functions

  • Awkward broadcasting

  • Reducing (sum/min/any/all)

  • Statistics (mean/var/std)

  • Using argmin/argmax

  • On GPUs

previous

How to ensure that an array is valid

next

How to perform computations with NumPy

By Jim Pivarski
© Copyright 2020 Jim Pivarski.