Overview of data loading methods v15

As with Oracle SQL*Loader, EDB*Loader supports three data loading methods:

  • Conventional path load
  • Direct path load
  • Parallel direct path load

Conventional path load is the default method used by EDB*Loader. Use basic insert processing to add rows to the table.

The advantage of a conventional path load is that table constraints and database objects defined on the table are enforced during a conventional path load. Table constraints and database objects include primary keys, not null constraints, check constraints, unique indexes, foreign key constraints, triggers, and so on.

One exception is that the EDB Postgres Advanced Server rules defined on the table aren't enforced. EDB*Loader can load tables on which rules are defined. However, the rules aren't executed. As a consequence, you can't load partitioned tables implemented using rules with EDB*Loader.

Note

Create EDB Postgres Advanced Server rules using the CREATE RULE command. EDB Postgres Advanced Server rules aren't the same database objects as rules and rule sets used in Oracle.

EDB*Loader also supports direct path loads. A direct path load is faster than a conventional path load but requires removing most types of constraints and triggers from the table. For more information, see Direct path load.

EDB*Loader supports parallel direct path loads. A parallel direct path load provides even greater performance improvement by permitting multiple EDB*Loader sessions to run simultaneously to load a single table. For more information, see Parallel direct path load.