[SPARK-11946][SQL] Audit pivot API for 1.6.

[rxin]

Currently pivot's signature looks like

@scala.annotation.varargs
defpivot(pivotColumn: Column, values: Column*):GroupedData@scala.annotation.varargs
defpivot(pivotColumn: String, values: Any*):GroupedData

I think we can remove the one that takes "Column" types, since callers should always be passing in literals. It'd also be more clear if the values are not varargs, but rather Seq or java.util.List.

I also made similar changes for Python.

[rxin]

cc @aray

[rxin]

@aray do you know why we have a "sort" in here?

[aray]

The sort is there to ensure that the output columns are in a consistent logical order.

[rxin]

ok thanks - i'm going to add a comment there to explain.

[rxin]

this is done as part of #9603 (comment) but it is way too small to deserve its own pr.

[SparkQA]

Test build #46591 has finished for PR 9929 at commit 73d37fc.

• This patch fails PySpark unit tests.
• This patch merges cleanly.
• This patch adds no public classes.

[davies]

Should we use _to_java_column(pivot_col) and _to_seq() here? or df.pivot(df.a) may fail

[SparkQA]

Test build #2104 has finished for PR 9929 at commit 73d37fc.

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

[davies]

LGTM

[rxin]

Thanks - merging this in.