[FLINK-32014][hotfix][javadoc] Fix incorrect javadoc regarding maxSpl…

…itMemorySize and add numSplits computation information (#14)
apache · May 10, 2023 · 6088c45 · 6088c45
1 parent f8ac343
commit 6088c45
Showing 1 changed file with 5 additions and 1 deletion.
diff --git a/...-cassandra/src/main/java/org/apache/flink/connector/cassandra/source/CassandraSource.java b/...-cassandra/src/main/java/org/apache/flink/connector/cassandra/source/CassandraSource.java
@@ -70,7 +70,7 @@
  *                   .build();
  *   }
  * };
- * long maxSplitMemorySize = ... //optional max split size in bytes. If not set, maxSplitMemorySize = tableSize / parallelism
+ * long maxSplitMemorySize = ... //optional max split size in bytes minimum is 10MB. If not set, maxSplitMemorySize = 64 MB
  * Source cassandraSource = new CassandraSource(clusterBuilder,
  *                                              maxSplitMemorySize,
  *                                              Pojo.class,
@@ -80,6 +80,10 @@
  * DataStream<Pojo> stream = env.fromSource(cassandraSource, WatermarkStrategy.noWatermarks(),
  * "CassandraSource");
  * }</pre>
+ *
+ * <p>Regarding performances, the source splits table data like this: numSplits =
+ * tableSize/maxSplitMemorySize. If tableSize cannot be determined or previous numSplits computation
+ * makes too few splits, falling back to numSplits=parallelism
  */
 @PublicEvolving
 public class CassandraSource<OUT>