hbase-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jim R. Wilson" <wilson.ji...@gmail.com>
Subject Re: Hbase query power
Date Tue, 20 May 2008 16:26:15 GMT
> Am I missing something ?

Nope.  Hbase is not an RDBMS.  The only thing which is indexed is the
rowkey, which is a string, and the only way to find things by any
other means is to scan for it.

If you want to be able to search on alternate keys, you can implement
this by using an additional table whose rowkeys are the indexes you'd
like to search on, then keep that up-to-date as new data is
entered/modified/removed from the main table.  You can also use a text
search engine like Lucene to maintain a dictionary of terms.

I've recently written an article to help people understand what
exactly Hbase is from a conceptual standpoint.  It may help you to
gague whether your existing dataset is suited for Hbase[1].

Good luck!

[1] http://jimbojw.com/wiki/index.php?title=Understanding_Hbase_and_BigTable

-- Jim R. Wilson (jimbojw)

On Tue, May 20, 2008 at 11:13 AM, shimon golan <sgolan@gmail.com> wrote:
> Hi all,
> I'm new to Hbase and consider using it as a replcaement to a RDBMS  such as
> MySQL due to increasing scalability and storage requirements.
> My main concern is whether the limited query language in HBase would not be
> an obstacle to migrate  - all I saw in the API where scanners that iterate
> by key ranges and simple getters which are far less expressive than SQL.
> Am I missing something ?
> Thanks in Advance,
> Shimon

View raw message