[SPARK-14350][SQL] EXPLAIN output should be in a single cell

[dongjoon-hyun]

What changes were proposed in this pull request?

EXPLAIN output should be in a single cell.

Before

scala> sql("explain select 1").collect()
res0: Array[org.apache.spark.sql.Row] = Array([== Physical Plan ==], [WholeStageCodegen], [:  +- Project [1 AS 1#1]], [:     +- INPUT], [+- Scan OneRowRelation[]])

After

scala> sql("explain select 1").collect()
res1: Array[org.apache.spark.sql.Row] =
Array([== Physical Plan ==
WholeStageCodegen
:  +- Project [1 AS 1#4]
:     +- INPUT
+- Scan OneRowRelation[]])

Or,

scala> sql("explain select 1").head
res1: org.apache.spark.sql.Row =
[== Physical Plan ==
WholeStageCodegen
:  +- Project [1 AS 1#5]
:     +- INPUT
+- Scan OneRowRelation[]]

Please note that Spark-shell(Scala-shell) trims long string output. So, you may need to use println to get full strings.

scala> println(sql("explain codegen select 'a' as a group by 1").head)
[Found 2 WholeStageCodegen subtrees.
== Subtree 1 / 2 ==
WholeStageCodegen
...
/* 059 */   }
/* 060 */ }

]

How was this patch tested?

Pass the Jenkins tests. (Testcases are updated.)

[hvanhovell]

LGTM

[dongjoon-hyun]

Thank you for fast review, @hvanhovell !

[SparkQA]

Test build #54802 has finished for PR 12137 at commit a1a9177.

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

[hvanhovell]

Merging to master. Thanks!

[dongjoon-hyun]

Oh, thank you! @hvanhovell . :)