Well, you can create an empty Hive table in Orc format and use --hive-override in sqoop
Alternatively you can use --hive-import and set hive.default.format I recommend to define the schema properly on the command line, because sqoop detection of formats is based on jdbc (Java) types which is not optimal. For example, a decimal(20,2) in oracle should be a decimal(20,2) in hive and not a double. This may also a good opportunity to review the database model in general. Especially for keys one should use a numeric type and not varchar. This saves a lot of space in the column and can be looked up much faster. > On 31 Jan 2016, at 14:15, Ashok Kumar <ashok34...@yahoo.com> wrote: > > Thanks, > > Can sqoop create this table as ORC in Hive? > > > On Sunday, 31 January 2016, 13:13, Ashok Kumar <ashok34...@yahoo.com> wrote: > > > Thanks. > > Can sqoop create this table as ORC in Hive? > > > On Sunday, 31 January 2016, 13:11, Nitin Pawar <nitinpawar...@gmail.com> > wrote: > > > check sqoop > > On Sun, Jan 31, 2016 at 6:36 PM, Ashok Kumar <ashok34...@yahoo.com> wrote: > Hi, > > What is the easiest method of importing data from an Oracle 11g table to Hive > please? This will be a weekly periodic job. The source table has 20 million > rows. > > I am running Hive 1.2.1 > > regards > > > > > > -- > Nitin Pawar > > > >
