tree: f4c59e2f912753b06f9706e870bb83aff36af72e [path history] [tgz]
  1. CMakeLists.txt
  3. src/

Python Wrapper for Datasketches


The easiest way to install the python wrapper is to run pip install git+

If you prefer to downlioad the source first, be sure to clone the repo with --recursive to ensure you get the python binding library (pybind11):

git clone --recursive
cd incubator-datasketches-cpp
pip install .

In the event you do not have pip installed, you can invoke the setup script directly by replacing the last line above with python3 install.


Having installed the library, loading the Datasketches library in Python is simple: from datasketches import *.

Available Sketch Classes

  • KLL
    • kll_ints_sketch
    • kll_floats_sketch
  • Frequent Items
    • frequent_strings_sketch
    • Error types are frequent_items_error_type.{NO_FALSE_NEGATIVES | NO_FALSE_POSITIVES}
  • Theta
    • update_theta_sketch
    • compact_theta_sketch (cannot be instantiated directly)
    • theta_union
    • theta_intersection
    • theta_a_not_b
  • HLL
    • hll_sketch
    • hll_union
    • Target HLL types are tgt_hll_type.{HLL_4 | HLL_6 | HLL_8}
  • CPC
    • cpc_sketch
    • cpc_union

Known Differences from C++

The Python API largely mirrors the C++ API, with a few minor exceptions: The primary known differences are that Python on modern platforms does not support unsigned integer values or numeric values with fewer than 64 bits. As a result, you may not be able to produce identical sketches from within Python as you can with Java and C++. Loading those sketches after they have been serialized from another language will work as expected.

We have also removed reliance on a builder class for theta sketches as Python allows named arguments to the constructor, not strictly positional arguments.