SPARK-3635
Find Strongly Connected Components with Graphx has a small bugSPARK-4060
MLlib, exposing special rdd functions to the public
Use -agentlib instead of -Xdebug in sbt-launch-lib.bash for debugging
Inconsistent spelling 'MLlib' and 'MLLib'
Change partitionStrategy to utilize PartitionStrategy.fromString(_) to match edgeStorageLevel and vertexStorageLevel syntax in Analytics.scala
Add .schemaRDD to JavaSchemaRDD
All columns of the application table in the history page are sorted in alphabetical order
rdd.setName returns None in PySpark
The example of building with sbt should be "sbt assembly" instead of "sbt compile"
Not suitable error message in spark-shell.cmd
typo comment in bin/utils.shSPARK-3838
Python code example for Word2Vec in user guide
MovieLensALS example fails from including Params in closure
Truncate appName in WebUI if it is too long
Cleanup scalastyle.txt at the end of running dev/scalastyle
Remove deprecation in jekyll for pygment in _config.yml
HiveQuerySuite SET tests depend on map orderings
fix redundant conditional
SPARK-2558
Mention --queue argument in YARN documentation
Typo in sbt/sbt- SPARK-4711
MLlib optimization: docs should suggest how to choose optimizer
Fix MLlib compilation warnings
SparkContext accepts local[0] as a master URL
Confusable variable name.
.gitignore in /python includes wrong directory.
Some scripts bin\*.cmd pollutes environment variables in Windows.
Kinesis examples set master as local
Add support for hadoop-2.5 + bump jets3t version
Log file name is hard coded even though there is a variable '$LOG_FILE '
Block generator throttling in NetworkReceiverSuite is flaky
Use scala.collection.Map for pattern matching instead of using Predef.Map (it is scala.collection.immutable.Map)
Jar url validation is not enough for Jar file
Closing shuffle writers we swallow more important exception
Some clean-up work after the refactoring of MLlib's SerDe for PySpark
Do not override user-defined conf_dir in spark-config.sh
Scala version is wrong in dev/audit-release/blank_sbt_build/build.sbt
Consolidate usages of <driver>
Do not set `spark.executor.instances` if not needed (yarn-cluster)
Display "Executor ID" for Tasks in Stage page
Sphinx build warnings
SparkSubmitDriverBootstrapper should return exit code of driver process
A problem of EdgePartitionBuilder in Graphx
FixedLengthBinaryRecordReader should readFully
remove redundant spark.executor.memory in doc
A corrupted format in Sphinx documents and building warnings
pyspark will not use ipython on Windows
let ApplicationMaster accept executor memory argument in same format as JVM memory strings
The version number of Spark in _config.yaml is wrong.
Yarn backend's default file replication should match HDFS's default one
[GraphX] add overrided count for EdgeRDD