Pandas 1.0 has been released. The popular data analysis and manipulation library for Python has become, since its first release in 2008, an industry standard adopted by data scientist as one of their main tools. It has undoubtedly contributed to elevate Python as the mainstream language for data science and analysis. However, its first official major version was released only yesterday, 12 years after its first minor version, and the documentation was made available on a newly redesigned website.

Among the new features that Pandas 1.0 brings are:

  • A new to_markdown() method to convert a data frame to Markdown. Bloggers like me are going to love this.
  • Custom windows for rolling operations: windows can now have a length defined by indexes instead of fixed.
  • The rolling apply() methods can now be run on Numba giving significant performance gains.

Several experimental features are also introduced, such as an NA data type for missing values (that users of R will recognize), and dedicated String and Boolean data types.

You can check the full list of enhanceents and new features on the official release page in the documentation.