Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Plain View
Accumulo, mail # user - Map Reduce over whole row


+
Mike Hugo 2013-01-11, 22:20
+
John Vines 2013-01-11, 22:24
Copy link to this message
-
Re: Map Reduce over whole row
Keith Turner 2013-01-11, 23:31
On Fri, Jan 11, 2013 at 5:24 PM, John Vines <[EMAIL PROTECTED]> wrote:
> The WholeRowIterator takes the entire row and encapsulates it in the same
> order it was scanned. Because everything is sorted, the columns in teh
> WholeRowIterator are maintained in the same order, and the PeekingIterator
> does not change this behavior. So you should be fine.

I do not think the WholeRowIterator is not used.   I think
AccumuloRowInputFormat uses o.a.a.c.client.RowIterator wrapping a
scanner.

The peeking iterator should give keys in sorted order.

>
> John
>
>
> On Fri, Jan 11, 2013 at 5:20 PM, Mike Hugo <[EMAIL PROTECTED]> wrote:
>>
>> I'm writing a map reduce job that utilizes the AccumuloRowInputFormat so I
>> can do some processing on an entire row.  When working with things in this
>> way, will PeekingIterator always give me the columns of the row in sorted
>> order?
>>
>> Thanks,
>>
>> Mike
>
>