[FEA] Support element_at with non-literal index

[viadea]

I wish we can support element_at with non-literal index.

For example:

import org.apache.spark.sql.types._
import org.apache.spark.sql.Row

val arrayData = Seq(
    Row("John",List("apple","orange","banana"),1),
    Row("David",List("apple","orange","banana"),2),
    Row("Harry",List("apple","other"),1)
)

val arraySchema = new StructType().add("name",StringType).add("fruits", ArrayType(StringType)).add("favorite",IntegerType)

val df = spark.createDataFrame(spark.sparkContext.parallelize(arrayData),arraySchema)
df.write.mode("overwrite").format("parquet").save("/tmp/testparquet")
val df2=spark.read.parquet("/tmp/testparquet")
df2.createOrReplaceTempView("df2")

spark.sql("select name,element_at(fruits,favorite) from df2").show()

The related unsupported messages:

!Expression <ElementAt> element_at(fruits#92, favorite#93, false) cannot run on GPU because ordinal only supports IntegerType if it is a literal value; index/key only supports IntegerType if it is a literal value

[revans2]

We should update GpuGetArrayItem while we are at it, to make both consistent.

[firestarman]

Depends on rapidsai/cudf#10340