## What changes were proposed in this pull request?

This PR proposes to both,

**Do not allow json arrays with multiple elements and return null in `from_json` with `StructType` as the schema.**

Currently, it only reads the single row when the input is a json array. So, the codes below:

```scala
import org.apache.spark.sql.functions._
import org.apache.spark.sql.types._
val schema = StructType(StructField("a", IntegerType) :: Nil)
Seq(("""[{"a": 1}, {"a": 2}]""")).toDF("struct").select(from_json(col("struct"), schema)).show()
```
prints

```
+--------------------+
|jsontostruct(struct)|
+--------------------+
|                 [1]|
+--------------------+
```

This PR simply suggests to print this as `null` if the schema is `StructType` and input is json array.with multiple elements

```
+--------------------+
|jsontostruct(struct)|
+--------------------+
|                null|
+--------------------+
```

**Support json arrays in `from_json` with `ArrayType` as the schema.**

```scala
import org.apache.spark.sql.functions._
import org.apache.spark.sql.types._
val schema = ArrayType(StructType(StructField("a", IntegerType) :: Nil))
Seq(("""[{"a": 1}, {"a": 2}]""")).toDF("array").select(from_json(col("array"), schema)).show()
```

prints

```
+-------------------+
|jsontostruct(array)|
+-------------------+
|         [[1], [2]]|
+-------------------+
```

## How was this patch tested?

Unit test in `JsonExpressionsSuite`, `JsonFunctionsSuite`, Python doctests and manual test.

Author: hyukjinkwon <gurwls223@gmail.com>

Closes #16929 from HyukjinKwon/disallow-array.