[
https://issues.apache.org/jira/browse/HBASE-15516?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15209092#comment-15209092
]
Ted Yu commented on HBASE-15516:
--------------------------------
Don't know why there were two hunks on the same source file in the patch.
{code}
[ERROR]
/Users/tyu/trunk/hbase-spark/src/main/scala/org/apache/hadoop/hbase/spark/HBaseContext.scala:452:
error: double definition:
[ERROR] method hbaseRDD:[U](tableName: org.apache.hadoop.hbase.TableName, scan:
org.apache.hadoop.hbase.client.Scan, f:
((org.apache.hadoop.hbase.io.ImmutableBytesWritable,
org.apache.hadoop.hbase.client.Result)) => Seq[U])(implicit evidence$6:
scala.reflect.ClassTag[U])org.apache.spark.rdd.RDD[U] and
[ERROR] method hbaseRDD:[U](tableName: org.apache.hadoop.hbase.TableName, scan:
org.apache.hadoop.hbase.client.Scan, f:
((org.apache.hadoop.hbase.io.ImmutableBytesWritable,
org.apache.hadoop.hbase.client.Result)) => U)(implicit evidence$5:
scala.reflect.ClassTag[U])org.apache.spark.rdd.RDD[U] at line 436
[ERROR] have same type after erasure: (tableName:
org.apache.hadoop.hbase.TableName, scan: org.apache.hadoop.hbase.client.Scan,
f: Function1, implicit evidence$6: reflect.ClassTag)org.apache.spark.rdd.RDD
[ERROR] def hbaseRDD[U: ClassTag](tableName: TableName, scan: Scan,
[ERROR] ^
{code}
Please fix the compilation error.
> Add flatMap to hbaseRDD
> -----------------------
>
> Key: HBASE-15516
> URL: https://issues.apache.org/jira/browse/HBASE-15516
> Project: HBase
> Issue Type: Improvement
> Components: spark
> Reporter: MahmoudHanafy
> Priority: Minor
> Attachments: HBASE-15516.patch
>
>
> HBaseContext supports reading RDDs using hbaseRDD method with map function, I
> think it will be good also to add flatMap to hbaseRDD.
> Another improvement:
> Currently reading RDDs is done by using method hbaseRDD with map function as
> the default method for loading RDDs. When trying to just load the RDD without
> mapping, you are also loading the RDD and map all elements to itself. So I
> think using hbaseRDD method without mapping function for loading RDDs would
> be better.
> Also, I can contribute to hbase-spark module. If there any issues related to
> this module, Please tell me about them
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
