accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Keith Turner (JIRA)" <j...@apache.org>
Subject [jira] [Comment Edited] (ACCUMULO-1337) Create Whole Column Family Iterator
Date Wed, 24 Apr 2013 18:33:16 GMT

    [ https://issues.apache.org/jira/browse/ACCUMULO-1337?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13640763#comment-13640763
] 

Keith Turner edited comment on ACCUMULO-1337 at 4/24/13 6:31 PM:
-----------------------------------------------------------------

Pushpinder, thanks for the patch.  Some unit test would be nice.  I do not think the continue
case is correct.   After an Accumulo client has received a batch of key values, it will reseek
with the last key in the batch non-inclusive.

So I am thinking the following code should not test the case 'sk.getColumnFamilyData().length()
== 0'.  The wholerowiterator always returns 0 len column families.   This iterator does not
do that.

{code:java}
    if (sk != null && sk.getColumnFamilyData().length() == 0 && sk.getColumnQualifierData().length()
== 0 && sk.getColumnVisibilityData().length() == 0
        && sk.getTimestamp() == Long.MAX_VALUE && !range.isStartKeyInclusive())
{
{code}

Could the WholeRowItertor be generalized?  Maybe create a more general version of the whole
row iterator, and have the whole row iterator extend that for backwards compatibility.  


                
      was (Author: kturner):
    Pushpinder, thanks for the patch.  Some unit test would be nice.  I do not think the continue
case is correct.   After an Accumulo has received a batch of key values, it will reseek with
the last key in the batch non-inclusive.

So I am thinking the following code should not test the case 'sk.getColumnFamilyData().length()
== 0'.  The wholerowiterator always returns 0 len column families.   This iterator does not
do that.

{code:java}
    if (sk != null && sk.getColumnFamilyData().length() == 0 && sk.getColumnQualifierData().length()
== 0 && sk.getColumnVisibilityData().length() == 0
        && sk.getTimestamp() == Long.MAX_VALUE && !range.isStartKeyInclusive())
{
{code}

Could the WholeRowItertor be generalized?  Maybe create a more general version of the whole
row iterator, and have the whole row iterator extend that for backwards compatibility.  


                  
> Create Whole Column Family Iterator
> -----------------------------------
>
>                 Key: ACCUMULO-1337
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-1337
>             Project: Accumulo
>          Issue Type: New Feature
>          Components: tserver
>    Affects Versions: 1.5.0, 1.4.3
>            Reporter: Pushpinder Heer
>            Assignee: Keith Turner
>            Priority: Trivial
>         Attachments: ACCUMULO-1337.patch
>
>
> Add similar functionality as WholeRowIterator, but group by row/cf.  WholeColumnFamilyIterator.

--
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

Mime
View raw message