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

Switch to Threaded View
HBase, mail # user - Coprocessor Aggregation supposed to be ~20x slower than Scans?


Copy link to this message
-
Re: Coprocessor Aggregation supposed to be ~20x slower than Scans?
Jimmy Xiang 2012-05-16, 17:28
Yes, it is fixed in CDH4.  It will be in the coming release.

Thanks,
Jimmy

On Tue, May 15, 2012 at 5:34 PM, Ted Yu <[EMAIL PROTECTED]> wrote:

> Hopefully this gets fixed in
>
> https://repository.cloudera.com/artifactory/public/org/apache/hbase/hbase/0.92.0-cdh4b2-SNAPSHOT/
>
> A developer from Cloudera would be able to better help you.
>
> On Tue, May 15, 2012 at 5:30 PM, anil gupta <[EMAIL PROTECTED]> wrote:
>
> > Hi Ted,
> >
> > I looked into hbase-0.92.0-cdh4b1-20120206.193413-23-sources.jar and it
> > also doesn't have it.
> >
> > On Tue, May 15, 2012 at 5:07 PM, Ted Yu <[EMAIL PROTECTED]> wrote:
> >
> > > Why did you need to decompile ?
> > > Here is the source code:
> > >
> > >
> > >
> >
> https://repository.cloudera.com/artifactory/public/org/apache/hbase/hbase/0.92.0-cdh4b1-SNAPSHOT/
> > >
> > > On Tue, May 15, 2012 at 4:58 PM, anil gupta <[EMAIL PROTECTED]>
> > wrote:
> > >
> > > > Hi Ted,
> > > >
> > > > I decompiled the hbase-0.92.0-cdh4b1.jar using JD-GUI and in
> > > > validateParameter method i don't find that condition.
> > > >
> > > > Thanks,
> > > > Anil
> > > >
> > > > On Tue, May 15, 2012 at 1:37 PM, Ted Yu <[EMAIL PROTECTED]> wrote:
> > > >
> > > > > I checked the code in Apache HBase 0.92 and trunk. I see the
> > following
> > > > line
> > > > > in validateParameters():
> > > > >            !Bytes.equals(scan.getStopRow(),
> > > HConstants.EMPTY_END_ROW))) {
> > > > >
> > > > > Can you confirm that the bug is in cdh4b1 only ?
> > > > >
> > > > > Sorry for not doing the validation earlier.
> > > > >
> > > > > On Tue, May 15, 2012 at 12:09 PM, anil gupta <
> [EMAIL PROTECTED]>
> > > > > wrote:
> > > > >
> > > > > > Oh i c.. Now if i look closely at your gmail id then i can see
> your
> > > > > name. I
> > > > > > was totally confused.
> > > > > >
> > > > > > So, you want to force the user to specify stopRow if the filter
> is
> > > not
> > > > > > used? What if the user just wants to scan the table from startRow
> > > till
> > > > > the
> > > > > > end of table? In your solution user will have explicitly set the
> > > > stopRow
> > > > > as
> > > > > > HConstants.EMPTY_END_ROW. Do we really want to force this?
> > > > > >
> > > > > > As per your solution the code would look like this:
> > > > > >      if(scan.hasFilter())
> > > > > >      {  if (scan == null || (Bytes.equals(scan.getStartRow(),
> > > > > > scan.getStopRow()) && !Bytes.equals(scan.getStartRow(),
> > > > > > HConstants.EMPTY_START_ROW)) ||
> > (Bytes.compareTo(scan.getStartRow(),
> > > > > > scan.getStopRow()) > 0 &&
> > > > > >             !Bytes.equals(scan.getStopRow(),
> > > HConstants.EMPTY_END_ROW)
> > > > > )) {
> > > > > >      throw new IOException(
> > > > > >          "Agg client Exception: Startrow should be smaller than
> > > > > Stoprow");
> > > > > >    } else if (scan.getFamilyMap().size() != 1) {
> > > > > >      throw new IOException("There must be only one family.");
> > > > > >    }
> > > > > >      }
> > > > > >      else
> > > > > >      {  if (scan == null || (Bytes.equals(scan.getStartRow(),
> > > > > > scan.getStopRow()) && !Bytes.equals(scan.getStartRow(),
> > > > > > HConstants.EMPTY_START_ROW)) ||
> Bytes.compareTo(scan.getStartRow(),
> > > > > > scan.getStopRow()) > 0) {
> > > > > >          throw new IOException(
> > > > > >               "Agg client Exception: Startrow should be smaller
> > than
> > > > > > Stoprow");
> > > > > >        } else if (scan.getFamilyMap().size() != 1) {
> > > > > >          throw new IOException("There must be only one family.");
> > > > > >        }
> > > > > >      }
> > > > > >
> > > > > > Let me know your thoughts.
> > > > > >
> > > > > > Thanks,
> > > > > > Anil
> > > > > >
> > > > > >
> > > > > > On Tue, May 15, 2012 at 11:46 AM, Ted Yu <[EMAIL PROTECTED]>
> > > wrote:
> > > > > >
> > > > > > > Anil:
> > > > > > > I am having trouble accessing JIRA.
> > > > > > >
> > > > > > > Ted Yu and Zhihong Yu are the same person :-)
> > > > > > >
> > > > > > > I think it would be good to remind user of aggregation client