[SPARK-5244] [SQL] add coalesce() in sql parser

[adrian-wang]

No description provided.

[SparkQA]

Test build #25530 has started for PR 4040 at commit 0ac8e8f.

• This patch merges cleanly.

[SparkQA]

Test build #25530 has finished for PR 4040 at commit 0ac8e8f.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[AmplabJenkins]

Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/25530/
Test PASSed.

[JoshRosen]

I've reverted this PR because it conflicted with another PR and has led to test compilation failures in SQLQuerySuite:

[info] Compiling 25 Scala sources to /private/tmp/spark/sql/catalyst/target/scala-2.10/test-classes...
[info] Compiling 38 Scala sources and 3 Java sources to /private/tmp/spark/sql/core/target/scala-2.10/test-classes...
[error] /private/tmp/spark/sql/core/src/test/scala/org/apache/spark/sql/SQLQuerySuite.scala:90: overloaded method value checkAnswer with alternatives:
[error]   (rdd: org.apache.spark.sql.SchemaRDD,expectedAnswer: org.apache.spark.sql.Row)Unit <and>
[error]   (rdd: org.apache.spark.sql.SchemaRDD,expectedAnswer: Seq[org.apache.spark.sql.Row])Unit
[error]  cannot be applied to (org.apache.spark.sql.SchemaRDD, Int)
[error]     checkAnswer(
[error]     ^
[error] /private/tmp/spark/sql/core/src/test/scala/org/apache/spark/sql/SQLQuerySuite.scala:93: ambiguous reference to overloaded definition,
[error] both method checkAnswer in class QueryTest of type (rdd: org.apache.spark.sql.SchemaRDD, expectedAnswer: org.apache.spark.sql.Row)Unit
[error] and  method checkAnswer in class QueryTest of type (rdd: org.apache.spark.sql.SchemaRDD, expectedAnswer: Seq[org.apache.spark.sql.Row])Unit
[error] match argument types (org.apache.spark.sql.SchemaRDD,Null)
[error]     checkAnswer(
[error]     ^
[error] two errors found
[error] (sql/test:compile) Compilation failed
[error] Total time: 29 s, completed Jan 21, 2015 2:24:18 PM

I'm reverting now to unblock pull request testing.

I believe that this may have been due to some recent refactoring in other parts of SQL, which Jenkins didn't detect because this wasn't retested recently.

[adrian-wang]

I have put code here together into #4057