Github user rdblue commented on a diff in the pull request:
https://github.com/apache/spark/pull/21145#discussion_r186498369
--- Diff:
sql/core/src/main/java/org/apache/spark/sql/sources/v2/reader/ReadTask.java ---
@@ -22,20 +22,20 @@
import org.apache.spark.annotation.InterfaceStability;
/**
- * A reader factory returned by {@link
DataSourceReader#createDataReaderFactories()} and is
+ * A read task returned by {@link DataSourceReader#createReadTasks()} and
is
--- End diff --
Now I'm rethinking the suggestion: `InputSplit` is a well-known Hadoop
class that we probably shouldn't duplicate. What about using `InputPartition`
instead? That makes it clear that the partitioning is on the input data and
uses the more common term in Spark.
Is everyone okay with this? @jose-torres @gengliangwang @cloud-fan @henryr
@arunmahadevan @gatorsmile?
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
