[
https://issues.apache.org/jira/browse/HBASE-17933?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15975161#comment-15975161
]
Yi Liang commented on HBASE-17933:
----------------------------------
Hi Sean,
Thanks for review, for your second comment, when I start designing the API, I
also want to use the approach you mentioned above. But since it is a java api,
we have some limitation in Java language. From its scala API
{code}
def bulkLoad[T](rdd:RDD[T],
tableName: TableName,
flatMap: (T) => Iterator[(KeyFamilyQualifier, Array[Byte])],
stagingDir:String,
familyHFileWriteOptionsMap:
util.Map[Array[Byte], FamilyHFileWriteOptions] =
new util.HashMap[Array[Byte], FamilyHFileWriteOptions],
compactionExclude: Boolean = false,
maxSize:Long = HConstants.DEFAULT_MAX_FILE_SIZE)
{code}
As you can see, the scala API allow callers to provide a method can transform
any type [T] to a key-value pair (KeyFamilyQualifier, Array[Byte]). But in
Java, a method(actually a class that implement mapFunction class) can not
directly transfer T to a kv pair like scala, since Java does not support (k,v)
pair I need to transfer it to a defined type, and in scala transfer this type
to the kv pair.
> [hbase-spark] Support Java api for bulkload
> --------------------------------------------
>
> Key: HBASE-17933
> URL: https://issues.apache.org/jira/browse/HBASE-17933
> Project: HBase
> Issue Type: New Feature
> Components: spark
> Affects Versions: 2.0.0
> Reporter: Yi Liang
> Assignee: Yi Liang
> Fix For: 2.0.0
>
> Attachments: HBase-17933-V1.patch
>
>
> In JavaHBaseContext, there are java api for bulkPut, bulkDelete ...., but no
> Java api for bulkload. And this jira will add bulkload java api to hbase-spark
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
