[beam] 02/03: Comment schema choices

[echauchot]

This is an automated email from the ASF dual-hosted git repository.

echauchot pushed a commit to branch spark-runner_structured-streaming
in repository https://gitbox.apache.org/repos/asf/beam.git

commit 1fe87911591940f49d2db95d3cda5a906a04fd96
Author: Etienne Chauchot <echauc...@apache.org>
AuthorDate: Tue Jan 15 13:35:52 2019 +0100

    Comment schema choices
---
 .../structuredstreaming/translation/batch/DatasetSourceBatch.java      | 3 +++
 1 file changed, 3 insertions(+)

diff --git 
a/runners/spark-structured-streaming/src/main/java/org/apache/beam/runners/spark/structuredstreaming/translation/batch/DatasetSourceBatch.java
 
b/runners/spark-structured-streaming/src/main/java/org/apache/beam/runners/spark/structuredstreaming/translation/batch/DatasetSourceBatch.java
index c4cfeaf..2a13d98 100644
--- 
a/runners/spark-structured-streaming/src/main/java/org/apache/beam/runners/spark/structuredstreaming/translation/batch/DatasetSourceBatch.java
+++ 
b/runners/spark-structured-streaming/src/main/java/org/apache/beam/runners/spark/structuredstreaming/translation/batch/DatasetSourceBatch.java
@@ -96,6 +96,9 @@ public class DatasetSourceBatch implements DataSourceV2, 
ReadSupport {
     @Override
     public StructType readSchema() {
       // TODO: find a way to extend schema with a WindowedValue schema
+      // we use a binary schema for now because:
+        // using a empty schema raises a indexOutOfBoundsException
+        // using a NullType schema stores null in the elements
       StructField[] array = new StructField[1];
       StructField binaryStructField = StructField
           .apply("binaryStructField", DataTypes.BinaryType, true, 
Metadata.empty());