[SPARK-21610][SQL] Corrupt records are not handled properly when creating a dataframe from a file

## What changes were proposed in this pull request?
```
echo '{"field": 1}
{"field": 2}
{"field": "3"}' >/tmp/sample.json
```

```scala
import org.apache.spark.sql.types._

val schema = new StructType()
  .add("field", ByteType)
  .add("_corrupt_record", StringType)

val file = "/tmp/sample.json"

val dfFromFile = spark.read.schema(schema).json(file)

scala> dfFromFile.show(false)
+-----+---------------+
|field|_corrupt_record|
+-----+---------------+
|1    |null           |
|2    |null           |
|null |{"field": "3"} |
+-----+---------------+

scala> dfFromFile.filter($"_corrupt_record".isNotNull).count()
res1: Long = 0

scala> dfFromFile.filter($"_corrupt_record".isNull).count()
res2: Long = 3
```
When the `requiredSchema` only contains `_corrupt_record`, the derived `actualSchema` is empty and the `_corrupt_record` are all null for all rows. This PR captures above situation and raise an exception with a reasonable workaround messag so that users can know what happened and how to fix the query.

## How was this patch tested?

Added test case.

Author: Jen-Ming Chung <jenmingisme@gmail.com>

Closes #18865 from jmchung/SPARK-21610.

docs/sql-programming-guide.md

@@ -1543,6 +1543,10 @@ options.
 
 # Migration Guide
 
+## Upgrading From Spark SQL 2.2 to 2.3
+
+  - Since Spark 2.3, the queries from raw JSON/CSV files are disallowed when the referenced columns only include the internal corrupt record column (named `_corrupt_record` by default). For example, `spark.read.schema(schema).json(file).filter($"_corrupt_record".isNotNull).count()` and `spark.read.schema(schema).json(file).select("_corrupt_record").show()`. Instead, you can cache or save the parsed results and then send the same query. For example, `val df = spark.read.schema(schema).json(file).cache()` and then `df.filter($"_corrupt_record".isNotNull).count()`.
+
 ## Upgrading From Spark SQL 2.1 to 2.2
 
   - Spark 2.1.1 introduced a new configuration key: `spark.sql.hive.caseSensitiveInferenceMode`. It had a default setting of `NEVER_INFER`, which kept behavior identical to 2.1.0. However, Spark 2.2.0 changes this setting's default value to `INFER_AND_SAVE` to restore compatibility with reading Hive metastore tables whose underlying file schema have mixed-case column names. With the `INFER_AND_SAVE` configuration value, on first access Spark will perform schema inference on any Hive metastore table for which it has not already saved an inferred schema. Note that schema inference can be a very time consuming operation for tables with thousands of partitions. If compatibility with mixed-case column names is not a concern, you can safely set `spark.sql.hive.caseSensitiveInferenceMode` to `NEVER_INFER` to avoid the initial overhead of schema inference. Note that with the new default `INFER_AND_SAVE` setting, the results of the schema inference are saved as a metastore key for future use. Therefore, the initial schema inference occurs only at a table's first access.

sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/json/JsonFileFormat.scala

@@ -113,6 +113,20 @@ class JsonFileFormat extends TextBasedFileFormat with DataSourceRegister {
       }
     }
 
+    if (requiredSchema.length == 1 &&
+      requiredSchema.head.name == parsedOptions.columnNameOfCorruptRecord) {
+      throw new AnalysisException(
+        "Since Spark 2.3, the queries from raw JSON/CSV files are disallowed when the\n" +
+        "referenced columns only include the internal corrupt record column\n" +
+        s"(named ${parsedOptions.columnNameOfCorruptRecord} by default). For example:\n" +
+        "spark.read.schema(schema).json(file).filter($\"_corrupt_record\".isNotNull).count()\n" +
+        "and spark.read.schema(schema).json(file).select(\"_corrupt_record\").show().\n" +
+        "Instead, you can cache or save the parsed results and then send the same query.\n" +
+        "For example, val df = spark.read.schema(schema).json(file).cache() and then\n" +
+        "df.filter($\"_corrupt_record\".isNotNull).count()."
+      )
+    }
+
     (file: PartitionedFile) => {
       val parser = new JacksonParser(actualSchema, parsedOptions)
       JsonDataSource(parsedOptions).readFile(

sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/json/JsonSuite.scala

@@ -2034,4 +2034,33 @@ class JsonSuite extends QueryTest with SharedSQLContext with TestJsonData {
       }
     }
   }
+
+  test("SPARK-21610: Corrupt records are not handled properly when creating a dataframe " +
+    "from a file") {
+    withTempPath { dir =>
+      val path = dir.getCanonicalPath
+      val data =
+        """{"field": 1}
+          |{"field": 2}
+          |{"field": "3"}""".stripMargin
+      Seq(data).toDF().repartition(1).write.text(path)
+      val schema = new StructType().add("field", ByteType).add("_corrupt_record", StringType)
+      // negative cases
+      val msg = intercept[AnalysisException] {
+        spark.read.schema(schema).json(path).select("_corrupt_record").collect()
+      }.getMessage
+      assert(msg.contains("only include the internal corrupt record column"))
+      intercept[catalyst.errors.TreeNodeException[_]] {
+        spark.read.schema(schema).json(path).filter($"_corrupt_record".isNotNull).count()
+      }
+      // workaround
+      val df = spark.read.schema(schema).json(path).cache()
+      assert(df.filter($"_corrupt_record".isNotNull).count() == 1)
+      assert(df.filter($"_corrupt_record".isNull).count() == 2)
+      checkAnswer(
+        df.select("_corrupt_record"),
+        Row(null) :: Row(null) :: Row("{\"field\": \"3\"}") :: Nil
+      )
+    }
+  }
 }