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

Switch to Plain View
Accumulo >> mail # user >> Organizing data in separate tables?

Roshan Punnoose 2013-01-18, 20:58
Copy link to this message
Re: Organizing data in separate tables?
"It depends."

There are features of a table that allow you to organize the data.  You can
offline and clone a whole table.  You can create locality groups.   You can
add/remove permissions on whole tables.  So, tables have great features.

Intermingling data in the same table has its benefits. For example, the
IntersectingIterator pattern allows you to distribute search, and perform
complex searches by grouping data elements with their indexes.

But, if you are just going to create a row prefix to separate sections that
would logically go into different tables, just put them in different tables.

On Fri, Jan 18, 2013 at 3:58 PM, Roshan Punnoose <[EMAIL PROTECTED]> wrote:

> Is it more performant to have multiple tables for different sets of data,
> or would having a properly split table into tablets have the same or better
> performance.
> I am storing data in three tables, but I was wondering if putting all that
> data into one table, with a row prefix, would have better performance.
> Thanks!
> Roshan Punnoose
David Medinets 2013-01-18, 22:46