test: Extend Spark Array functions: array_repeat , shuffle and slice test coverage

[erenavsarogullari]

Which issue does this PR close?

• Closes Extend Spark Array functions: array_repeat , shuffle and slice test coverage #20419.

Rationale for this change

This PR adds new positive test cases for datafusion-spark array functions: array_repeat , shuffle, slice for the following use-cases:

- nested function execution,
- different datatypes such as timestamp,
- casting before function execution

Also, being updated contributor-guide testing documentation with minor addition.

What changes are included in this PR?

Being added new positive test cases to datafusion-spark array functions: array_repeat , shuffle, slice.

Are these changes tested?

Yes, adding new positive test cases.

Are there any user-facing changes?

No

[getChan]

LGTM.

[comphead]

this looks unrelated to the PR

[Jefffrey]

It seems a nice improvement regardless

[erenavsarogullari]

I have covered this doc update here because of a minor change

[comphead]

not related to the PR, but how it comes shuffle which is non deterministic https://spark.apache.org/docs/latest/api/sql/#shuffle has stable output 🤔

[Jefffrey]

We're providing a seed for deterministic output

[erenavsarogullari]

In general, shuffle function is non-deterministic. However, it can also return deterministic result by passing seed for specially this kind of cases (e.g: test verification).

For reference:

• Spark shuffle function accepts optional randomSeed argument to :
  https://github.com/apache/spark/blob/master/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/collectionOperations.scala#L1264

• On the other hand, current DF-Spark shuffle func implementation support following both cases by accepting both arguments for the alignment with Spark:

shuffle(array)
shuffle(array, seed)

[Jefffrey]

We're providing a seed for deterministic output

[Jefffrey]

It seems a nice improvement regardless

[comphead]

Thanks everyone