hbase-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Mingtao Zhang <mail2ming...@gmail.com>
Subject Re: Scan result sequence
Date Sat, 16 Aug 2014 21:55:48 GMT
Hi Ted,

I used murmurhash. Actually I don't care about the sequence between the
group of a.com and b.com record. I am 120% :) as far as hash(a.com) comes
together with the timestamp sequence. (hash(b.com) could come either before
or after)

Best Regards,
Mingtao

On Sat, Aug 16, 2014 at 5:44 PM, Ted Yu <yuzhihong@gmail.com> wrote:

> How do you generate hash based on pageurl ?
> The order between hash(a.com) and hash(b.com <http://a.com/>) may not be
> what you expected.
>
> BTW See http://hbase.apache.org/book.html#row and
> http://hbase.apache.org/book.html#dm.sort
>
> Cheers
>
>
> On Sat, Aug 16, 2014 at 2:26 PM, Mingtao Zhang <mail2mingtao@gmail.com>
> wrote:
>
> > Hi,
> >
> > My rowkey is
> >
> > sessionid|hash(pageurl)|timestamp
> >
> > When I scan using a prefix filter with a specific sessionid, will it give
> > me the result in sequence? For example, the expected sequence in my mind
> > is:
> >
> > session1|hash(a.com)|1
> > session1|hash(a.com)|2
> > session1|hash(a.com)|3
> > session1|hash(b.com)|2.5
> > session1|hash(b.com)|5
> > session1|hash(b.com)|6
> > session1|hash(c.com)|3.5
> > session1|hash(c.com)|5.5
> > session1|hash(c.com)|7
> >
> > Thanks in advance!
> >
> > Best Regards,
> > Mingtao
> >
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message