If you use INSERT INTO to insert data into a Greenplum table from a table on the Hadoop file system that was defined as an external table with the gphdfs protocol, the data is copied in parallel.

For example:

INSERT INTO gpdb_table (select * from hdfs_ext_table);