[ https://issues.apache.org/jira/browse/HBASE-3737?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13465160#comment-13465160 ]
Lars Hofhansl commented on HBASE-3737: -------------------------------------- This is still the case in trunk. Also looking at the Delete(List<Delete>) code, the passed lists gets modified and will contain those Deletes that failed to be executed. The client presumably has to check and retry. I doubt anybody is doing that. Put(List<Put>) is similar (but worse IMHO). The call to the Put method happily returns even when there are left over Puts in the write buffer. > HTable - delete(List<Delete>) doesn't use writebuffer > ----------------------------------------------------- > > Key: HBASE-3737 > URL: https://issues.apache.org/jira/browse/HBASE-3737 > Project: HBase > Issue Type: Improvement > Reporter: Doug Meil > Priority: Minor > > I just realized that htable.delete(List<Delete>) doesn't use the writebuffer > and processes the list immediately, but htable.put(List<Put>) does use the > writebuffer (i.e., send when filled). Likewise, htable.delete(Delete) sends > immediately. > > Out of sheer curiosity, why? With the 'batch' methods now in place, it seems > like it would be consistent for 'delete' and 'put' to use the writebuffer > (assuming it is expanded to hold more than Puts), whereas 'batch' methods > process immediately. > This isn't a huge issue, but it does seem a little inconsistent. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira