Author: Daoyuan Wang <daoyuan.wang@intel.com>

This patch had conflicts when merged, resolved by
Committer: Reynold Xin <rxin@databricks.com>

Closes #6718 from adrian-wang/udflog2 and squashes the following commits:

3909f48 [Daoyuan Wang] math function: log2

Author: Cheng Hao <hao.cheng@intel.com>

Closes #6724 from chenghao-intel/length and squashes the following commits:

aaa3c31 [Cheng Hao] revert the additional change
97148a9 [Cheng Hao] remove the codegen testing temporally
ae08003 [Cheng Hao] update the comments
1eb1fd1 [Cheng Hao] simplify the code as commented
3e92d32 [Cheng Hao] use the selectExpr in unit test intead of SQLQuery
3c729aa [Cheng Hao] fix bug for constant null value in codegen
3641f06 [Cheng Hao] keep the length() method for registered function
8e30171 [Cheng Hao] update the code as comment
db604ae [Cheng Hao] Add code gen support
548d2ef [Cheng Hao] register the length()
09a0738 [Cheng Hao] add length support

Currently we only support `Seq[Expression]`, we should handle cases like `Seq[Seq[Expression]]` so that we can remove the unnecessary `GroupExpression`.

Author: Wenchen Fan <cloud0fan@outlook.com>

Closes #6706 from cloud-fan/clean and squashes the following commits:

60a1193 [Wenchen Fan] support nested expression sequence and remove GroupExpression

    case cs  CombineSum(expr) =>
        val calcType = expr.dataType
          expr.dataType match {
            case DecimalType.Fixed(_, _) =>
              DecimalType.Unlimited
            case _ =>
              expr.dataType
          }
calcType is always expr.dataType. credits are all belong to IntelliJ

Author: navis.ryu <navis@apache.org>

Closes #6736 from navis/SPARK-8285 and squashes the following commits:

20382c1 [navis.ryu] [SPARK-8285] [SQL] CombineSum should be calculated as unlimited decimal first

This PR change to use Long as internal type for TimestampType for efficiency, which means it will the precision below 100ns.

Author: Davies Liu <davies@databricks.com>

Closes #6733 from davies/timestamp and squashes the following commits:

d9565fa [Davies Liu] remove print
65cf2f1 [Davies Liu] fix Timestamp in SparkR
86fecfb [Davies Liu] disable two timestamp tests
8f77ee0 [Davies Liu] fix scala style
246ee74 [Davies Liu] address comments
309d2e1 [Davies Liu] use Long for TimestampType in SQL

Test cases for both StreamingLinearRegression and StreamingLogisticRegression, and code fix.

Edit:
This contribution is my original work and I license the work to the project under the project's open source license.

Author: Paavo <pparkkin@gmail.com>

Closes #6713 from pparkkin/streamingmodel-empty-rdd and squashes the following commits:

ff5cd78 [Paavo] Update strings to use interpolation.
db234cf [Paavo] Use !rdd.isEmpty.
54ad89e [Paavo] Test case for empty stream.
393e36f [Paavo] Ignore empty RDDs.
0bfc365 [Paavo] Test case for empty stream.

Set preferred locations for reduce tasks.
The basic design is that we maintain a map from reducerId to a list of (sizes, locations) for each
shuffle. We then set the preferred locations to be any machines that have 20% of more of the output
that needs to be read by the reduce task.  This will result in at most 5 preferred locations for
each reduce task.

Selecting the preferred locations involves O(# map tasks * # reduce tasks) computation, so we
restrict this feature to cases where we have fewer than 1000 map tasks and 1000 reduce tasks.

Author: Shivaram Venkataraman <shivaram@cs.berkeley.edu>

Closes #6652 from shivaram/reduce-locations and squashes the following commits:

492e25e [Shivaram Venkataraman] Remove unused import
2ef2d39 [Shivaram Venkataraman] Address code review comments
897a914 [Shivaram Venkataraman] Remove unused hash map
f5be578 [Shivaram Venkataraman] Use fraction of map outputs to determine locations Also removes caching of preferred locations to make the API cleaner
68bc29e [Shivaram Venkataraman] Fix line length
1090b58 [Shivaram Venkataraman] Change flag name
77ce7d8 [Shivaram Venkataraman] Merge branch 'master' of https://github.com/apache/spark into reduce-locations
e5d56bd [Shivaram Venkataraman] Add flag to turn off locality for shuffle deps
6cfae98 [Shivaram Venkataraman] Filter out zero blocks, rename variables
9d5831a [Shivaram Venkataraman] Address some more comments
8e31266 [Shivaram Venkataraman] Fix style
0df3180 [Shivaram Venkataraman] Address code review comments
e7d5449 [Shivaram Venkataraman] Fix merge issues
ad7cb53 [Shivaram Venkataraman] Merge branch 'master' of https://github.com/apache/spark into reduce-locations
df14cee [Shivaram Venkataraman] Merge branch 'master' of https://github.com/apache/spark into reduce-locations
5093aea [Shivaram Venkataraman] Merge branch 'master' of https://github.com/apache/spark into reduce-locations
0171d3c [Shivaram Venkataraman] Merge branch 'master' of https://github.com/apache/spark into reduce-locations
bc4dfd6 [Shivaram Venkataraman] Merge branch 'master' of https://github.com/apache/spark into reduce-locations
774751b [Shivaram Venkataraman] Fix bug introduced by line length adjustment
34d0283 [Shivaram Venkataraman] Fix style issues
3b464b7 [Shivaram Venkataraman] Set preferred locations for reduce tasks This is another attempt at #1697 addressing some of the earlier concerns. This adds a couple of thresholds based on number map and reduce tasks beyond which we don't use preferred locations for reduce tasks.

In client mode, if yarn was shut down with spark application running, the application will hang up after several retries(default: 30) because the exception throwed by YarnClientImpl could not be caught by upper level, we should exit in case that user can not be aware that.

The exception we wanna catch is [here](https://github.com/apache/hadoop/blob/branch-2.7.0/hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/io/retry/RetryInvocationHandler.java#L122), and I try to fix it refer to [MR](https://github.com/apache/hadoop/blob/branch-2.7.0/hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-jobclient/src/main/java/org/apache/hadoop/mapred/ClientServiceDelegate.java#L320).

Author: WangTaoTheTonic <wangtao111@huawei.com>

Closes #6717 from WangTaoTheTonic/SPARK-8273 and squashes the following commits:

28752d6 [WangTaoTheTonic] catch the throwed exception

SPARK_JAVA_OPTS was missed in reconstructing the launcher part, we should add it back so process launched by spark-class could read it properly. And so does `SPARK_DRIVER_MEMORY`.

The missing part is [here](https://github.com/apache/spark/blob/1c30afdf94b27e1ad65df0735575306e65d148a1/bin/spark-class#L97).

Author: WangTaoTheTonic <wangtao111@huawei.com>
Author: Tao Wang <wangtao111@huawei.com>

Closes #6741 from WangTaoTheTonic/SPARK-8290 and squashes the following commits:

bd89f0f [Tao Wang] make sure the memory setting is right too
e313520 [WangTaoTheTonic] spark class command builder need read SPARK_JAVA_OPTS

1. Add `log4j-defaults-repl.properties` that has log level WARN.
2. When logging is initialized, check whether inside the REPL. If so, use `log4j-defaults-repl.properties`.
3. Print the following information if using `log4j-defaults-repl.properties`:
```
Using Spark's repl log4j profile: org/apache/spark/log4j-defaults-repl.properties
To adjust logging level use sc.setLogLevel("INFO")
```

Author: zsxwing <zsxwing@gmail.com>

Closes #6734 from zsxwing/log4j-repl and squashes the following commits:

3835eff [zsxwing] Change default log level to WARN in the REPL

The root cause of SPARK-7527 is `createNullValue` returns an incompatible value `Byte(0)` for `char` and `boolean`.

This PR fixes it and corrects the class name of the main class, and also adds an unit test to demonstrate it.

Author: zsxwing <zsxwing@gmail.com>

Closes #6735 from zsxwing/SPARK-7527 and squashes the following commits:

bbdb271 [zsxwing] Use pattern match in createNullValue
b0a0e7e [zsxwing] Remove the noisy in the test output
903e269 [zsxwing] Remove the code for Utils.isInInterpreter == false
5f92dc1 [zsxwing] Fix createNullValue to return the correct null values and REPL mode detection

IBM Java has an extra method when we do getStackTrace(): this is "getStackTraceImpl", a native method. This causes two tests to fail within "DStreamScopeSuite" when running with IBM Java. Instead of "map" or "filter" being the method names found, "getStackTrace" is returned. This commit addresses such an issue by using dropWhile. Given that our current method is withScope, we look for the next method that isn't ours: we don't care about methods that come before us in the stack trace: e.g. getStackTrace (regardless of how many levels this might go).

IBM:
java.lang.Thread.getStackTraceImpl(Native Method)
java.lang.Thread.getStackTrace(Thread.java:1117)
org.apache.spark.rdd.RDDOperationScope$.withScope(RDDOperationScope.scala:104)

Oracle:
PRINTING STACKTRACE!!!
java.lang.Thread.getStackTrace(Thread.java:1552)
org.apache.spark.rdd.RDDOperationScope$.withScope(RDDOperationScope.scala:106)

I've tested this with Oracle and IBM Java, no side effects for other tests introduced.

Author: Adam Roberts <aroberts@uk.ibm.com>
Author: a-roberts <aroberts@uk.ibm.com>

Closes #6740 from a-roberts/RDDScopeStackCrawlFix and squashes the following commits:

13ce390 [Adam Roberts] Ensure consistency with String equality checking
a4fc0e0 [a-roberts] Update RDDOperationScope.scala

Read number of threads for RBackend from configuration.

[SPARK-8282] #comment Linking with JIRA

Author: Hossein <hossein@databricks.com>

Closes #6730 from falaki/SPARK-8282 and squashes the following commits:

33b3d98 [Hossein] Documented new config parameter
70f2a9c [Hossein] Fixing import
ec44225 [Hossein] Read number of threads for RBackend from configuration

The bug description is a little misleading: the actual issue is that
.py files are not handled correctly when distributed by YARN. They're
added to "spark.submit.pyFiles", which, when processed by context.py,
explicitly whitelists certain extensions (see PACKAGE_EXTENSIONS),
and that does not include .py files.

On top of that, archives were not handled at all! They made it to the
driver's python path, but never made it to executors, since the mechanism
used to propagate their location (spark.submit.pyFiles) only works on
the driver side.

So, instead, ignore "spark.submit.pyFiles" and just build PYTHONPATH
correctly for both driver and executors. Individual .py files are
placed in a subdirectory of the container's local dir in the cluster,
which is then added to the python path. Archives are added directly.

The change, as a side effect, ends up solving the symptom described
in the bug. The issue was not that the files were not being distributed,
but that they were never made visible to the python application
running under Spark.

Also included is a proper unit test for running python on YARN, which
broke in several different ways with the previous code.

A short walk around of the changes:
- SparkSubmit does not try to be smart about how YARN handles python
  files anymore. It just passes down the configs to the YARN client
  code.
- The YARN client distributes python files and archives differently,
  placing the files in a subdirectory.
- The YARN client now sets PYTHONPATH for the processes it launches;
  to properly handle different locations, it uses YARN's support for
  embedding env variables, so to avoid YARN expanding those at the
  wrong time, SparkConf is now propagated to the AM using a conf file
  instead of command line options.
- Because the Client initialization code is a maze of implicit
  dependencies, some code needed to be moved around to make sure
  all needed state was available when the code ran.
- The pyspark tests in YarnClusterSuite now actually distribute and try
  to use both a python file and an archive containing a different python
  module. Also added a yarn-client tests for completeness.
- I cleaned up some of the code around distributing files to YARN, to
  avoid adding more copied & pasted code to handle the new files being
  distributed.

Author: Marcelo Vanzin <vanzin@cloudera.com>

Closes #6360 from vanzin/SPARK-5479 and squashes the following commits:

bcaf7e6 [Marcelo Vanzin] Feedback.
c47501f [Marcelo Vanzin] Fix yarn-client mode.
46b1d0c [Marcelo Vanzin] Merge branch 'master' into SPARK-5479
c743778 [Marcelo Vanzin] Only pyspark cares about python archives.
c8e5a82 [Marcelo Vanzin] Actually run pyspark in client mode.
705571d [Marcelo Vanzin] Move some code to the YARN module.
1dd4d0c [Marcelo Vanzin] Review feedback.
71ee736 [Marcelo Vanzin] Merge branch 'master' into SPARK-5479
220358b [Marcelo Vanzin] Scalastyle.
cdbb990 [Marcelo Vanzin] Merge branch 'master' into SPARK-5479
7fe3cd4 [Marcelo Vanzin] No need to distribute primary file to executors.
09045f1 [Marcelo Vanzin] Style.
943cbf4 [Marcelo Vanzin] [SPARK-5479] [yarn] Handle --py-files correctly in YARN.

Author: Cheng Lian <lian@databricks.com>

Closes #6749 from liancheng/java-sample-fix and squashes the following commits:

5b44585 [Cheng Lian] Fixes a minor Java example error in SQL programming guide

Changed ```SparkEnv.actorSystem``` to be a function such that we can use the deprecated flag with it and added a deprecated message.

Author: Ilya Ganelin <ilya.ganelin@capitalone.com>

Closes #6731 from ilganeli/SPARK-7996 and squashes the following commits:

be43817 [Ilya Ganelin] Restored to val
9ed89e7 [Ilya Ganelin] Added a version info for deprecation
9610b08 [Ilya Ganelin] Converted actorSystem to function and added deprecated flag

Author: Daoyuan Wang <daoyuan.wang@intel.com>

Closes #6716 from adrian-wang/epi and squashes the following commits:

e2e8dbd [Daoyuan Wang] move tests
11b351c [Daoyuan Wang] add tests and remove pu
db331c9 [Daoyuan Wang] py style
599ddd8 [Daoyuan Wang] add py
e6783ef [Daoyuan Wang] register function
82d426e [Daoyuan Wang] add function entry
dbf3ab5 [Daoyuan Wang] add PI and E

This is a followup to #6712.

Author: Reynold Xin <rxin@databricks.com>

Closes #6739 from rxin/6712-followup and squashes the following commits:

fd9acfb [Reynold Xin] [SPARK-7886] Added unit test for HAVING aggregate pushdown.

This builds on #6710 and also uses FunctionRegistry for function lookup in HiveContext.

Author: Reynold Xin <rxin@databricks.com>

Closes #6712 from rxin/udf-registry-hive and squashes the following commits:

f4c2df0 [Reynold Xin] Fixed style violation.
0bd4127 [Reynold Xin] Fixed Python UDFs.
f9a0378 [Reynold Xin] Disable one more test.
5609494 [Reynold Xin] Disable some failing tests.
4efea20 [Reynold Xin] Don't check children resolved for UDF resolution.
2ebe549 [Reynold Xin] Removed more hardcoded functions.
aadce78 [Reynold Xin] [SPARK-7886] Use FunctionRegistry for built-in expressions in HiveContext.

Just replaced mutable.HashMap to ConcurrentHashMap

Author: navis.ryu <navis@apache.org>

Closes #6699 from navis/SPARK-7792 and squashes the following commits:

f03654a [navis.ryu] [SPARK-7792] [SQL] HiveContext registerTempTable not thread safe

This provides preliminary documentation pointing out how to use the
Hadoop free builds. I am hoping over time this list can grow to
include most of the popular Hadoop distributions.

Getting more people using these builds will help us long term reduce
the number of binaries we build.

Author: Patrick Wendell <patrick@databricks.com>

Closes #6729 from pwendell/hadoop-provided and squashes the following commits:

1113b76 [Patrick Wendell] [SPARK-6511] [Documentation] Explain how to use Hadoop provided builds

Just as a safeguard against DOM rewriting.

Author: Andrew Or <andrew@databricks.com>

Closes #6732 from andrewor14/dag-viz-trim and squashes the following commits:

7e9bacb [Andrew Or] [MINOR] [UI] DAG visualization: trim whitespace from input

There is a mistake in the URLs of the Scala section of FP-Growth in the MLlib Frequent Pattern Mining documentation. The URL points to https://spark.apache.org/docs/latest/api/java/org/apache/spark/mllib/fpm/FPGrowth.html which is the Java's API, the link should point to the Scala API https://spark.apache.org/docs/latest/api/scala/index.html#org.apache.spark.mllib.fpm.FPGrowth

There's another mistake in the FP-GrowthModel in the same section, the link points, again, to the Java's API https://spark.apache.org/docs/latest/api/java/org/apache/spark/mllib/fpm/FPGrowthModel.html, the link should point to the Scala API https://spark.apache.org/docs/latest/api/scala/index.html#org.apache.spark.mllib.fpm.FPGrowthModel

Author: FavioVazquez <favio.vazquezp@gmail.com>

Closes #6722 from FavioVazquez/fix-wrog-urls-mllib-fpgrowth and squashes the following commits:

e1ca54d [FavioVazquez] - Fixed wrong URLs in MLlib Frequent Pattern Mining, FP-Growth Scala section
ad882a3 [FavioVazquez] Merge remote-tracking branch 'upstream/master'
f27a20b [FavioVazquez] Merge remote-tracking branch 'upstream/master'
9af7074 [FavioVazquez] Merge remote-tracking branch 'upstream/master'
edab1ef [FavioVazquez] Merge remote-tracking branch 'upstream/master'
b2e2f8c [FavioVazquez] Merge remote-tracking branch 'upstream/master'

Author: MechCoder <manojkumarsivaraj334@gmail.com>

Closes #6720 from MechCoder/empty_model_check and squashes the following commits:

3a07de5 [MechCoder] Remove construct to get weights in StreamingLinearAlgorithm

I noticed there is a duplicated description about WAL.

```
To ensure zero-data loss, you have to additionally enable Write Ahead Logs in Spark Streaming. To ensure zero data loss, enable the Write Ahead Logs (introduced in Spark 1.2).
```

Let's remove the duplication.

I don't file this issue in JIRA because it's minor.

Author: Kousuke Saruta <sarutak@oss.nttdata.co.jp>

Closes #6719 from sarutak/remove-multiple-description and squashes the following commits:

cc9bb21 [Kousuke Saruta] Removed duplicated description about WAL

This patch switches to using FunctionRegistry for built-in expressions. It is based on #6463, but with some work to simplify it along with unit tests.

TODOs for future pull requests:
- Use static registration so we don't need to register all functions every time we start a new SQLContext
- Switch to using this in HiveContext

Author: Reynold Xin <rxin@databricks.com>
Author: Santiago M. Mola <santi@mola.io>

Closes #6710 from rxin/udf-registry and squashes the following commits:

6930822 [Reynold Xin] Fixed Python test.
b802c9a [Reynold Xin] Made UDF case insensitive.
e60d815 [Reynold Xin] Made UDF case insensitive.
852f9c0 [Reynold Xin] Fixed style violation.
e76a3c1 [Reynold Xin] Fixed parser.
52ddaba [Reynold Xin] Fixed compilation.
ee7854f [Reynold Xin] Improved error reporting.
ff906f2 [Reynold Xin] More robust constructor calling.
77b46f1 [Reynold Xin] Simplified the code.
2a2a149 [Reynold Xin] Merge pull request #6463 from smola/SPARK-7886
8616924 [Santiago M. Mola] [SPARK-7886] Add built-in expressions to FunctionRegistry.

Update to Netty 4.0.28-Final

Author: Sean Owen <sowen@cloudera.com>

Closes #6701 from srowen/SPARK-8101 and squashes the following commits:

f3b6369 [Sean Owen] Update to Netty 4.0.28-Final

JIRA: https://issues.apache.org/jira/browse/SPARK-7990

Author: Liang-Chi Hsieh <viirya@gmail.com>

Closes #6616 from viirya/multi_keys_equi_join and squashes the following commits:

cd5c888 [Liang-Chi Hsieh] Import reduce in python3.
c43722c [Liang-Chi Hsieh] For comments.
0400e89 [Liang-Chi Hsieh] Fix scala style.
cc90015 [Liang-Chi Hsieh] Add methods to facilitate equi-join on multiple joining keys.

Author: hqzizania <qian.huang@intel.com>

Closes #6190 from hqzizania/R and squashes the following commits:

1641f9e [hqzizania] fixes and add test units
bb3411a [hqzizania] Convert NAs to null type in SparkR DataFrames

This makes the constructor callable in Python. dbtsai

Author: Xiangrui Meng <meng@databricks.com>

Closes #6709 from mengxr/SPARK-8168 and squashes the following commits:

f871de4 [Xiangrui Meng] Add Python friendly constructor to PipelineModel

This was caused by this commit: f2713478

This patch does not attempt to fix the root cause of why the `VisibleForTesting` annotation causes a NPE in the shell. We should find a way to fix that separately.

Author: Andrew Or <andrew@databricks.com>

Closes #6711 from andrewor14/fix-spark-shell and squashes the following commits:

bf62ecc [Andrew Or] Prevent NPE in spark-shell

Use DoubleType instead to be more stable and robust.

Author: Reynold Xin <rxin@databricks.com>

Closes #6692 from rxin/SPARK-8148 and squashes the following commits:

6742ecc [Reynold Xin] [SPARK-8148] Do not use FloatType in partition column inference.

We already have a rule to do type coercion for fixed decimal and unlimited decimal in `WidenTypes`, so we don't need to handle them in `DecimalPrecision`.

Author: Wenchen Fan <cloud0fan@outlook.com>

Closes #6698 from cloud-fan/fix and squashes the following commits:

413ad4a [Wenchen Fan] remove duplicated cases

For Hadoop 1.x, `TaskAttemptContext` constructor clones the `Configuration` argument, thus configurations done in `HadoopFsRelation.prepareForWriteJob()` are not populated to *driver* side `TaskAttemptContext` (executor side configurations are properly populated). Currently this should only affect Parquet output committer class configuration.

Author: Cheng Lian <lian@databricks.com>

Closes #6669 from liancheng/spark-8121 and squashes the following commits:

73819e8 [Cheng Lian] Minor logging fix
fce089c [Cheng Lian] Adds more logging
b6f78a6 [Cheng Lian] Fixes compilation error introduced while rebasing
963a1aa [Cheng Lian] Addresses @yhuai's comment
c3a0b1a [Cheng Lian] Fixes InsertIntoHadoopFsRelation job initialization

1. explicitly import implicit conversion support.
2. use .nonEmpty instead of .size > 0
3. use val instead of var
4. comment indention

Author: Daoyuan Wang <daoyuan.wang@intel.com>

Closes #6700 from adrian-wang/shimsimprove and squashes the following commits:

d22e108 [Daoyuan Wang] several fix for HiveShim

Author: Daoyuan Wang <daoyuan.wang@intel.com>

Closes #6434 from adrian-wang/joinerr and squashes the following commits:

ee1b64f [Daoyuan Wang] break line
f7c53e9 [Daoyuan Wang] to IllegalArgumentException
f8dea2d [Daoyuan Wang] sys.err to IllegalStateException
be82259 [Daoyuan Wang] change new exception to sys.err

Return null if catch EOFException, just like function "asKeyValueIterator" in this class

Author: Mingfei <mingfei.shi@intel.com>

Closes #6703 from shimingfei/returnNull and squashes the following commits:

205deec [Mingfei] return null if catch EOFException

1. Prevent creating a map of data to find numFeatures
2. If model is empty, then initialize with a zero vector of numFeature

Author: MechCoder <manojkumarsivaraj334@gmail.com>

Closes #6684 from MechCoder/spark-8140 and squashes the following commits:

7fbf5f9 [MechCoder] [SPARK-8140] Remove empty model check in StreamingLinearAlgorithm And other minor cosmits