Oracle 10g Data Pump

1) Dump file are create on the server side.
2) Data moving is much faster, with the use of Parallelism.
3) No need of commit, it is taken care by datapump
4) Buffer is no longer required to be specified, and also commit is done automatically
5) We could stop/start the data pump job.


We have to create directory for storing the dumpfile

SQL>create direcoty dp as '~/path/to dir/'

Grant permision to users on above, if they are required to use it

SQL>grant read,write on db to <user>;

Now lets export

expdp scott/tiger dumpfile='dp:exp.dmp' logfile=dp:exp.log tables=employee,dept job_name=tables_dump



Datapump Parameters

remap_schema=source_schema:target_schema
content=all|data_only|metadata
exclude=triggers                 //Include  everything except trigger
include=constraint               //include just constraint
query=dept:"where deptno=20"
sqlfile=dir:file                 //DDL statement for the items you are exporitng
network_link=<database_link>


impdp two/two2 DIRECTORY=dp NETWORK_LINK=src tables=TWO.TWOEMP remap_schema=two:three PARALLEL=4 LOGFILE=impdp.log