Extending Pandas using Apache Arrow and Numba

With the latest release of Pandas the ability to extend it with custom dtypes was introduced. Using Apache Arrow as the in-memory storage and Numba for fast, vectorized computations on these memory regions, it is possible to extend Pandas in pure Python while achieving the same performance of the built-in types. In the talk we implement a native string type as an example.

Transcript

1. 1 PyData Berlin 2018 Uwe L. Korn Extending Pandas using Apache Arrow and Numba
2. 2 PyData Berlin 2018 Uwe L. Korn Extending Pandas using Apache Arrow and Numba
3. 3 PyData Berlin 2018 Uwe L. Korn Strings, Strings, please give me Strings!
4. 4 • Senior Data Scientist at Blue Yonder (@BlueYonderTech) • Apache {Arrow, Parquet} PMC • Data Engineer and Architect with heavy focus around Pandas About me xhochy mail@uwekorn.com
5. 5 1. Shortcomings of Pandas 2. ExtensionArrays 3. Arrow for storage 4. Numba for compute 5. All the stuff Agenda
6. 6 Pandas Series • Payload stored in a numpy.ndarray • Index for data alignment • Rich analytical API • Accessors like .dt or .str
7. 7 Shortcomings • Limited to NumPy data types, otherwise object • NumPy’s focus is numerical data and tensors • Pandas performs well when NumPy performs well • Most popular: • no native variable-length strings • integers are non-nullable
8. 8 What’s the problem?
9. 9 What’s the problem?
10. 10 Why are objects bad? Python Data Science Handbook, Jake VanderPlas; O’Reilly Media, Nov 2016 https://jakevdp.github.io/PythonDataScienceHandbook/02.01-understanding-data-types.html
11. 11 Extending Pandas (0.23+) • Two new interfaces: • ExtensionDtype • What type of scalars? • ExtensionArray • Implement basic array ops • Pandas provides algorithms on top
12. 10x !!1 12
13. 13 Extending Pandas (0.23+) • _from_sequence • _from_factorized • __getitem__ • __len__ • dtype • nbytes • isna • copy • _concat_same_type https://pandas.pydata.org/pandas-docs/stable/generated/pandas.api.extensions.ExtensionArray.html 13
14. 14 Apache Arrow • Specification for in-memory columnar data layout • No overhead for cross-system communication • Designed for efficiency (exploit SIMD, cache locality, ..) • Exchange data without conversion between Python, C++, C(glib), Ruby, Lua, R, JavaScript, Go, Rust, Matlab and the JVM • Brought Parquet to Pandas and made PySpark fast (@pandas_udf)
15. 15 Nice properties • More native datatypes: string, date, nullable int, list of X, … • Everything is nullable • Memory can be chunked • Zero-copy to other ecosystems like Java / R • Highly efficient I/O
16. 16 Not so nice properties • Still a young project • Not much analytic on top (yet!) • Core is in modern C++ • Extremely fast but hard to extend in Python
17. 17 Writing Algorithms in Python is easy! but slow
18. 18 Photo by Matthew Brodeur on Unsplash
19. 19 Fast for-loops with Numba
20. 20 Anatomy of an Arrow StringArray • 3 memory buffers • bitmap to indicate valid (non-null) entries • uint32 array of offsets: „where does the string start“ • uint8 array of characters (UTF-8 encoded) • int64 offset • allows zero-copy slicing
21. 21 Numba @jitclass
22. 22 Numba @jitclass
23. 23 Photo by Niklas Tidbury on Unsplash
24. 24 Fletcher https://github.com/xhochy/fletcher • Implements Extension{Array,Dtype} with Apache Arrow as storage • Uses Numba to implement the necessary analytic on top
25. Demo 25
26. 26 Fletcher Demo
27. 27 Fletcher Demo
28. 28 Fletcher Demo
29. 29 Fletcher Demo
30. 30 ExtensionArray Implementations https://github.com/ContinuumIO/cyberpandas IPArray (PR) https://github.com/geopandas/geopandas GeometryArray (WIP) https://github.com/xhochy/fletcher Apache Arrow + Numba backed Arrays
31. 31 Photo by Israel Sundseth on Unsplash pip install fletcher
32. 32 By JOEXX (Own work) [CC BY-SA 3.0 (http://creativecommons.org/licenses/by-sa/3.0)], via Wikimedia Commons By JOEXX (Own work) [CC BY-SA 3.0 (http://creativecommons.org/licenses/by-sa/3.0)], via Wikimedia Commons 24. - 26. October + 2 days of sprints (27/28.10.) ZKM Karlsruhe, DE Karlsruhe Call for Participation opens next week.
33. 33 I’m Uwe Korn Twitter: @xhochy https://github.com/xhochy Thank you!