[
https://issues.apache.org/jira/browse/ACCUMULO-1801?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13970204#comment-13970204
]
Russ Weeks commented on ACCUMULO-1801:
--------------------------------------
Just wanted to add a bit of unexpected behaviour that came down to this issue:
when the RowDeletingIterator is applied to a table, clients who call
o.a.a.core.client.ScannerBase.fetchColumn prior to scanning must remember to
also fetch the empty column family+empty column qualifier. Otherwise, "deleted"
rows will still continue to show up until the table is compacted. Very
confusing behaviour! It makes sense once you understand what's going on but
it's pretty counter-intuitive.
> Create a SelectColumnsIterator
> ------------------------------
>
> Key: ACCUMULO-1801
> URL: https://issues.apache.org/jira/browse/ACCUMULO-1801
> Project: Accumulo
> Issue Type: Improvement
> Components: client
> Reporter: Mike Drob
> Priority: Minor
> Fix For: 1.7.0
>
>
> A user can already limit columns on a scanner, however there is not an easy
> way to limit the columns propagated in the middle of an iterator stack. For
> example, if a user has data like the following:
> {noformat}
> user-1 age:25
> user-1 occupation:Student
> user-2 age:45
> user-2 occupation:Engineer
> {noformat}
> and wants to get only the users that are younger than 30, it is easy to write
> a custom {{RowFilter}} to do that.
> However, to get only the occupation of users under 30, there would need to be
> additional filtering done on the client side. If the rows get large or
> particularly numerous, then this becomes undesirable. And setting the fetch
> columns is not an option since that would hide the age column from the
> {{RowFilter}}.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
