Doc changes for new nested JSON reader [skip ci]

docs/compatibility.md

@@ -296,38 +296,21 @@ The JSON format read is a very experimental feature which is expected to have so
 it by default. If you would like to test it, you need to enable `spark.rapids.sql.format.json.enabled` and 
 `spark.rapids.sql.format.json.read.enabled`.
 
-Currently, the GPU accelerated JSON reader doesn't support column pruning, which will likely make 
-this difficult to use or even test. The user must specify the full schema or just let Spark infer 
-the schema from the JSON file. eg,
-
-We have a `people.json` file with below content
-
+Reading input containing invalid JSON format (in any row) will throw runtime exception.
+An example of valid input is as following:
 ``` console
-{"name":"Michael"}
 {"name":"Andy", "age":30}
 {"name":"Justin", "age":19}
 ```
 
-Both below ways will work
-
-- Inferring the schema
-
-  ``` scala
-  val df = spark.read.json("people.json")
-  ```
-
-- Specifying the full schema
-
-  ``` scala
-  val schema = StructType(Seq(StructField("name", StringType), StructField("age", IntegerType)))
-  val df = spark.read.schema(schema).json("people.json")
-  ```
-
-While the below code will not work in the current version,
+The following input is invalid and will cause error:
+```console
+{"name":"Andy", "age":30} ,,,,
+{"name":"Justin", "age":19}
+```
 
-``` scala
-val schema = StructType(Seq(StructField("name", StringType)))
-val df = spark.read.schema(schema).json("people.json")
+```console
+{"name":  Justin", "age":19}
 ```
 
 ### JSON supporting types