[
https://issues.apache.org/jira/browse/HBASE-16646?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15503283#comment-15503283
]
Ashish Singhi commented on HBASE-16646:
---------------------------------------
bq. public int run(String dirPath, Map<byte[], List<Path>> map, TableName
tableName) throws Exception{
map is not being used. You ned to call the new doBulkLoad api written.
{quote} public void prepareHFileQueue(Map<byte[], List<Path>> map, Table table,
Deque<LoadQueueItem> queue, boolean validateHFile, boolean silence)
throws IOException {
{quote}
validateHFile parameter is not used, can be removed.
Can you also add test case covering this scenario.
> Enhance LoadIncrementalHFiles API to accept store file paths as input
> ---------------------------------------------------------------------
>
> Key: HBASE-16646
> URL: https://issues.apache.org/jira/browse/HBASE-16646
> Project: HBase
> Issue Type: Improvement
> Reporter: Ted Yu
> Assignee: Ted Yu
> Attachments: 16646.v0.txt, 16646.v1.txt, 16646.v2.txt
>
>
> Currently LoadIncrementalHFiles takes the directory (output path) as input
> parameter.
> In some scenarios (incremental restore of bulk loaded hfiles), the List of
> paths to hfiles is known.
> LoadIncrementalHFiles can take the List as input parameter and proceed with
> loading.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)