[SPARK-49155][SQL][SS] Use more appropriate parameter type to construct `GenericArrayData`

[LuciferYang]

What changes were proposed in this pull request?

Referring to the test results of GenericArrayDataBenchmark, using an Array of Any to construct GenericArrayData is more efficient compared to other scenarios:

https://github.com/apache/spark/blob/master/sql/catalyst/benchmarks/GenericArrayDataBenchmark-results.txt

OpenJDK 64-Bit Server VM 17.0.11+9-LTS on Linux 6.5.0-1018-azure
AMD EPYC 7763 64-Core Processor
constructor:                              Best Time(ms)   Avg Time(ms)   Stdev(ms)    Rate(M/s)   Per Row(ns)   Relative
------------------------------------------------------------------------------------------------------------------------
arrayOfAny                                            6              6           0       1620.1           0.6       1.0X
arrayOfAnyAsObject                                    6              6           0       1620.1           0.6       1.0X
arrayOfAnyAsSeq                                     155            155           1         64.7          15.5       0.0X
arrayOfInt                                          253            254           1         39.6          25.3       0.0X
arrayOfIntAsObject                                  252            253           1         39.7          25.2       0.0X

So this pr optimizes some processes of constructing GenericArrayData in Spark code:

1. In ArraysZip#eval and XPathList#nullSafeEval, the originally defined arrays of specific types are changed to data of type AnyRef to avoid additional collection copying when constructing GenericArrayData. This is because the Array[AnyRef] type can also match the case array: Array[Any] => array branch in the following code:

https://github.com/apache/spark/blob/af70aafd330fdbb6ce0d5b3efbcb180cda488695/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/util/GenericArrayData.scala#L42-L48

2. In HistogramNumeric#eval, an IndexedSeq[InternalRow] was originally used to construct GenericArrayData. Since the length of the collection is known, it can be refactored to use Array[AnyRef] to construct GenericArrayData.

3. For other cases, when constructing GenericArrayData, the current input parameter is ${input}.toArray now. It is changed to ${input}.toArray[Any] to avoid another collection copy during the construction of GenericArrayData.

Why are the changes needed?

Using an Array of Any|AnyRef to construct GenericArrayData can improve performance by reducing collection copying.

Does this PR introduce any user-facing change?

No

How was this patch tested?

Pass GitHub Actions

Was this patch authored or co-authored using generative AI tooling?

No