**Thanks to the many contributors to this release!**
release-0.5.0
**We *strongly* encourage all users to use this latest release because of the bug fix described below!**
**CRITICAL bug fix**
* All previous versions of GraphFrames had a bug which can affect ConnectedComponents and other algorithms.
* This fixes a bug in indexing vertices with non-Integer IDs.
* It can affect all algorithms which are wrappers around GraphX, including ConnectedComponents, PageRank, and others.
* The bug surfaces when the input DataFrame is non-deterministic. E.g., running an algorithm on a DataFrame just loaded from disk should be fine in previous releases, but running that algorithm on a DataFrame produced using shuffling, unions, and other operators can cause incorrect results. This issue is fixed in this release.
**New features**
* Python API for aggregateMessages for building custom graph algorithms
* Scala API for parallel personalized PageRank, wrapping the GraphX implementation. This is only available when using GraphFrames with Spark 2.1+.