accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Chris McCubbin (JIRA)" <j...@apache.org>
Subject [jira] [Created] (ACCUMULO-1836) mutation could take key-value pairs in put() to be more efficient
Date Thu, 31 Oct 2013 19:15:20 GMT
Chris McCubbin created ACCUMULO-1836:
----------------------------------------

             Summary: mutation could take key-value pairs in put() to be more efficient
                 Key: ACCUMULO-1836
                 URL: https://issues.apache.org/jira/browse/ACCUMULO-1836
             Project: Accumulo
          Issue Type: Improvement
            Reporter: Chris McCubbin
            Priority: Minor


Right now, if I have a key-value pair and I want to make an insert out of it, I have to do
something like this:

{code}
m.put(key.getColumnFamily(), key.getColumnQualifier(), new ColumnVisibility(key.getColumnVisibility()),
value);
{code}

However, I believe this results in the data in the key being array copied twice which is pretty
inefficient (profiling shows this to be a fairly significant portion of our running time on
heavy ingest). Since I think exposing the byte arrays directly is not preferred, perhaps there
could be a method Mutation.put(Key,Value) that efficiently re-uses the byte arrays. Maybe
it assumes the rows are the same, or maybe it could have an option for checking that.



--
This message was sent by Atlassian JIRA
(v6.1#6144)

Mime
View raw message