Github user lw-lin commented on a diff in the pull request:
https://github.com/apache/spark/pull/13575#discussion_r66466672
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/json/JsonFileFormat.scala
---
@@ -146,16 +173,53 @@ class JsonFileFormat extends FileFormat with
DataSourceRegister {
}
}
- override def toString: String = "JSON"
-
override def hashCode(): Int = getClass.hashCode()
override def equals(other: Any): Boolean =
other.isInstanceOf[JsonFileFormat]
}
-private[json] class JsonOutputWriter(
- path: String,
- bucketId: Option[Int],
+/**
+ * A factory for generating [[OutputWriter]]s for writing json files. This
is implemented different
+ * from the 'batch' JsonOutputWriter as this does not use any
[[OutputCommitter]]. It simply
+ * writes the data to the path used to generate the output writer. Callers
of this factory
+ * has to ensure which files are to be considered as committed.
+ */
+private[json] class StreamingJsonOutputWriterFactory(
+ sqlConf: SQLConf,
+ dataSchema: StructType,
+ hadoopConf: Configuration,
+ options: Map[String, String]) extends StreamingOutputWriterFactory {
+
+ private val serializableConf = {
+ val conf = Job.getInstance(hadoopConf).getConfiguration
+ JsonFileFormat.prepareConfForWriting(conf, options)
+ new SerializableConfiguration(conf)
+ }
+
+ /**
+ * Returns a [[OutputWriter]] that writes data to the give path without
using an
+ * [[OutputCommitter]].
+ */
+ override private[sql] def newWriter(path: String): OutputWriter = {
+ val hadoopTaskAttempId = new TaskAttemptID(new TaskID(new JobID,
TaskType.MAP, 0), 0)
+ val hadoopAttemptContext =
+ new TaskAttemptContextImpl(serializableConf.value,
hadoopTaskAttempId)
+ // Returns a 'streaming' JsonOutputWriter
+ new JsonOutputWriterBase(dataSchema, hadoopAttemptContext) {
+ override private[json] val recordWriter: RecordWriter[NullWritable,
Text] =
+ createNoCommitterTextRecordWriter(
+ path,
+ hadoopAttemptContext,
+ (c: TaskAttemptContext, ext: String) => { new
Path(s"$path.json$ext") })
+ }
+ }
+}
+
+/**
+ * Base JsonOutputWriter class for 'batch' JsonOutputWriter and
'streaming' JsonOutputWriter. The
+ * writing logic to a single file resides in this base class.
+ */
+private[json] abstract class JsonOutputWriterBase(
--- End diff --
This `JsonOutputWriterBase` is basically the original `JsonOutputWriter`
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
