HBase-1512 is about implementing aggregation by coprocessor.
https://issues.apache.org/jira/browse/HBASE-1512
But it still needs to scan each region to get the row count. It's like a
light-weight map/red method, so it's not the super ``quick'' way.
Jonathan mentioned to use a separate table to record/count rows in
another email. I read something similar before at:
http://whynosql.com/aggregation-with-hbase/
FYI.
-Mingjie
On 12/02/2010 05:57 PM, Michael Segel wrote:
Uhm just a sick thought... with the talks about coprocessors...
Couldn't one write something to monitor the WAL and manage record/row counts?
Sort of use coprocessors to manage statistics and such.
Ok... I'll go back to my carb deprived diet now... :-)
-Mike
Date: Thu, 2 Dec 2010 16:52:09 -0800
Subject: Re: Quick way to find the number of rows in a Table
From: [email protected]
To: [email protected]
What Michael said.
Run the rowcounter M/R job:
Type:
./bin/hadoop jar hbase.jar
... to learn more, read
http://hbase.apache.org/docs/r0.89.20100924/apidocs/org/apache/hadoop/hbase/mapreduce/package-summary.html#package_description
St.Ack
On Thu, Dec 2, 2010 at 4:10 PM, Michael Segel<[email protected]> wrote:
The quickest way is a m/r job.
From: [email protected]
To: [email protected]
Date: Thu, 2 Dec 2010 16:02:44 -0800
Subject: Quick way to find the number of rows in a Table
Hi
I am going to have a process running that will add rows to a table. The row
id is just a counter of the number of rows inserted. Is there a quick way to
programmatically find out the number of rows in a table so when I start adding
rows the count isn't disrupted? Also, is there a better way to assign a row Id?
Thanks
-Pete
