[
https://issues.apache.org/jira/browse/PHOENIX-1069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14056488#comment-14056488
]
Jeffrey Zhong commented on PHOENIX-1069:
----------------------------------------
[~jaywong] The HFileOutputFormat2 is in hbase 0.98 &
trunk(https://github.com/apache/hbase/blob/0.98/hbase-server/src/main/java/org/apache/hadoop/hbase/mapreduce/HFileOutputFormat2.java).
{quote}
it seems like if the table has an index, we should just import it automatically
rather than have the user specify the index on the command line. Otherwise,
after the bulk load, it'll be in an invalid state (i.e. the index will be empty
while the table will have data), right?
{quote}
[~giacomotaylor] Yes, you're right. By default, all indexes are built during
loading. The --index-table optional command switch is only for situation when
an index table creation failed and we want to just reload this particular
index.
> Improve CsvBulkLoadTool to build indexes when loading data.
> -----------------------------------------------------------
>
> Key: PHOENIX-1069
> URL: https://issues.apache.org/jira/browse/PHOENIX-1069
> Project: Phoenix
> Issue Type: Improvement
> Affects Versions: 3.0.0, 4.0.0
> Reporter: Jeffrey Zhong
> Attachments: phoenix-1069.patch
>
>
> Currently CsvBulkLoadTool only imports data not the indexes. It will be
> convenient for people to load data & indexes at same time. Here come the
> JIRA.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
