[
https://issues.apache.org/jira/browse/ACCUMULO-2178?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13890765#comment-13890765
]
Josh Elser commented on ACCUMULO-2178:
--------------------------------------
Definitely, [~sonixbp]. For some set of Ranges, we would want to have some
specification of sequential column ranges that we {{seek()}} to, linearly scan
over, then move to the beginning of that next column in this row.
This is definitely a bit trickier than it sounds, especially when dealing with
the sparse nature of columns in Accumulo. Defining exactly how these column
ranges would work up front would be good and help us do the efficient thing at
that tabletserver instead of blindly filtering every row.
> ColumnFamilyRangeIterator
> -------------------------
>
> Key: ACCUMULO-2178
> URL: https://issues.apache.org/jira/browse/ACCUMULO-2178
> Project: Accumulo
> Issue Type: Improvement
> Reporter: Corey J. Nolet
> Priority: Minor
>
> It would be extremely useful in situations where you would like to specify a
> range of rows to iterate and also a range of column families within that
> range of rows. The current API requires that you specify a start/end key for
> each row that you will be scanning in the range of column families.
> There would be a lot gained from the ability to first seek to the beginning
> column specified in the range of columns, return all keys/values inside of
> that range, and then seek to the beginning of the range for the following row
> when the ending column family in the current row has been reached.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
