Customizing Anaconda

z/OS® IzODA Anaconda (FMID HANA110) contains the Anaconda data science ecosystem, which consists of the Anaconda package manager and a library of packages that data scientists and application developers can assemble into specific stacks of runtime capabilities to support particular data science applications.

Anaconda is primarily a Python-based environment favored by data scientists who are familiar with the analytics capabilities that have been built around the Python language. This environment complements and overlaps some of the capabilities that are available in z/OS IzODA Spark (FMID HSPK120). Together, Anaconda and z/OS Spark provide a complete data science platform that can support a broad range of analytics applications.

A more complete description of the Anaconda environment is available at Anaconda.

Note: Users of Spark 2.2.0 and earlier who are migrating to a newer level of Python should continue to use Python 3.6.

Anaconda package management

Anaconda package management is provided through the conda command line interface. This is used to manage the root environment of installed packages. A base set of packages is provided with the Anaconda component, and a repository of additional packages is available through the IzODA channel hosted in the Anaconda Cloud. For a complete list of available packages, see Package Repository for IzODA.

Anaconda repositories

Anaconda repositories are dynamic and flexible. The conda package manager can be used to install or uninstall any version of a package whenever it is published to a repository. This capability allows Anaconda to be framework for managing all types of functional packages, and makes it ideal for handling runtime environments based on open source.

Detailed information about all of the capabilities of the Anaconda ecosystem on z/OS, as well as the installation and configuration of the z/OS IzODA Anaconda component, is available on github at Anaconda Overview.