New in 2.0a4¶
- The
sercore
extension, together with most of thepydoop.mapreduce
subpackage, has been rewritten from scratch. Now it’s simpler and slightly faster (much faster when using a combiner)- Opaque splits are now auto-deserialized to
context.input_split.payload
JobConf
is now fully compatible withdict
- Compilation of avro-parquet-based examples is now much faster
- The Hadoop simulator has been dropped
- Bug fixes and performance improvements
New in 2.0a3¶
- Support for Hadoop 3
- Support for opaque binary input splits
- Moved terasort example to https://github.com/crs4/pydoop-examples
New in 2.0a2¶
- Support for Amazon EMR
New in 2.0a1¶
- Added support for Python 3
- Dropped support for Hadoop 1
- Dropped old MapReduce API
- Dropped JPype HDFS backend
- Added Terasort example
- Bug fixes and performance improvements