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

Switch to Threaded View
Hive >> mail # dev >> Hive projects for Google Summer of code 2012 ?


Copy link to this message
-
Re: Hive projects for Google Summer of code 2012 ?
Ed,

Did you see this?

http://google-melange.appspot.com/gsoc/project/google/gsoc2012/bharathv/29002

On Wed, Mar 7, 2012 at 12:15 PM, Edward Capriolo <[EMAIL PROTECTED]>wrote:

> For the record we are 0-2 in gsoc.
>
> On Wed, Mar 7, 2012 at 1:42 AM, Namit Jain <[EMAIL PROTECTED]> wrote:
> > Done
> >
> >
> > On 3/6/12 10:32 PM, "bharath vissapragada"
> > <[EMAIL PROTECTED]> wrote:
> >
> >>Hi Namit,
> >>
> >>Is it possible to add Hive-1362 [1] to the list? I am interested in
> >>working
> >>on that and I've contacted Ashutosh regarding that.
> >>
> >>Thanks and Regards,
> >>
> >>
> >>[1]  https://issues.apache.org/jira/browse/HIVE-1362
> >>
> >>On Wed, Mar 7, 2012 at 6:46 AM, Namit Jain <[EMAIL PROTECTED]> wrote:
> >>
> >>>
> >>>
> https://issues.apache.org/jira/secure/IssueNavigator.jspa?requestId=12319
> >>>27
> >>> 0
> >>>
> >>>
> >>>
> >>> 0 was in the next line by mistake
> >>> Look for HIVE jiras.
> >>>
> >>>
> >>>
> >>> On 3/6/12 7:37 PM, "Alexis De La Cruz Toledo" <[EMAIL PROTECTED]>
> >>>wrote:
> >>>
> >>> >Sorry Mr. Namit Jain I can't see the jiras.
> >>> >
> >>> >Thanks
> >>> >
> >>> >El 6 de marzo de 2012 19:41, Namit Jain <[EMAIL PROTECTED]> escribió:
> >>> >
> >>> >> I filed a couple of gsoc 2012 jiras.
> >>> >>
> >>> >> Please submit your proposal if you are interested.
> >>> >>
> >>> >>
> >>> >>
> >>>
> >>>
> https://issues.apache.org/jira/secure/IssueNavigator.jspa?requestId=12319
> >>> >>27
> >>> >> 0
> >>> >>
> >>> >>
> >>> >>
> >>> >>
> >>> >> Thanks,
> >>> >> -namit
> >>> >>
> >>> >>
> >>> >> On 2/7/12 6:32 PM, "Alexis De La Cruz Toledo" <[EMAIL PROTECTED]>
> >>> >>wrote:
> >>> >>
> >>> >> >Hi Namit Jain, I'm Alexis, I'm a master student, I'm studying in
> >>> >> >Cinvestav,
> >>> >> >df Mexico.
> >>> >> >I'm interesting in collaborating with Hive. I'd like to do my
> thesis
> >>> >>work
> >>> >> >in something about Hive.
> >>> >> >I'd like to participate in Google Summer of Code 2012 too.
> >>> >> >
> >>> >> >The issues that you propose I find interesting, particularly in the
> >>> >>next
> >>> >> >issues:
> >>> >> >* The topic 2. Indexed Joins.
> >>> >> >* [PO] Optimize Joins using Bloom Filters of this page
> >>> >> >https://cwiki.apache.org/confluence/display/Hive/Roadmap.
> >>> >> >
> >>> >> >Can you tell me something more about it?
> >>> >> >What is problem to be solved?
> >>> >> >What benefits we hope to gain?
> >>> >> >
> >>> >> >This because I want to raise my thesis problem.
> >>> >> >By another hand, can you be my mentor in Google Summer of Code, if
> I
> >>> >>work
> >>> >> >with this topics?
> >>> >> >
> >>> >> >Thanks.
> >>> >> >
> >>> >> >El 5 de febrero de 2012 19:58, Namit Jain <[EMAIL PROTECTED]> escribió:
> >>> >> >
> >>> >> >> Hi Alexis/Bharath,
> >>> >> >>
> >>> >> >> Great to see your interests. If you looking for ideas, some
> things
> >>> >>that
> >>> >> >> will be very useful are:
> >>> >> >>
> >>> >> >> 1. Removing the map-join hint completely.
> >>> >> >>   Moving all processing to runtime.
> >>> >> >>   Currently, bucketed map joins and sort merge joins are
> >>>completely
> >>> >> >> driven off hints.
> >>> >> >>   It would be very helpful to the community, and also clean up
> the
> >>> >>code
> >>> >> >>a
> >>> >> >> lot.
> >>> >> >>
> >>> >> >> 2. Indexed Joins.
> >>> >> >>   Something that would be really useful -
> >>> >> >>   If the basic infrastructure is ready, it can eventually be used
> >>>to
> >>> >> >>join
> >>> >> >> tables
> >>> >> >>   stored outside also (say Hbase).
> >>> >> >>
> >>> >> >> 3. Metastore understanding hierarchy.
> >>> >> >>   For eg: if a table is partitioned by ds and hr,
> >>> >> >>   A valid partition on ds does not exist. This is a very common
> >>> >>usecase
> >>> >> >> on many
> >>> >> >>   applications, and the current work-around is to have signal
> >>>tables
> >>> >>for
> >>> >> >> ds
> >>> >> >>   un-necessarily complicating the process.
> >>> >> >>
> >>> >> >>
> >>> >> >> If you are interested, I would be happy to provide more details.