source and destination tables have same schema. Are you suggesting to use sort by any of the partitions ?
On Mon, Nov 2, 2015 at 6:52 PM, Jörn Franke <[email protected]> wrote: > What is the create table statement? You may want to insert everything into > the orc table (sorted on x and/or y) and then apply the where statement in > your queries on the orc table. > > > On 02 Nov 2015, at 13:36, Kashif Hussain <[email protected]> wrote: > > > > Hi, > > I am trying to insert data into orc table from a text table.The query is > as follows : > > > > insert into table test_orc(x,y) select * from test where x=4 and y=5 ; > > > > The MR is taking a lot of time and all but one reducer got completed > quickly.But the last reducer is taking a lot of time. > > The total data size is 2.9 TB. I am not able to understand why only one > reducer is taking so much of time ? Is it because whole map output went to > a single reducer ? > > > > > > Regards, > > Kashif > > >
