Oracle¶
To be able to use the Oracle API to perform bulk data loads, first follow these steps:
On the Oracle database, enable direct path loads. To prepare it for direct path loads, the administrator of Oracle needs to run the setup script
catldr.sql
of Oracle, which creates the necessary views in Oracle to do direct loads. The administrator only needs to run this script once for each Oracle installation you plan to do direct loads to.The documentation of Oracle explains how direct path loads work.
Important
If you cannot enable this feature on the database, do not continue with these steps. If you enable bulk data loads in Virtual DataPort and the feature of Oracle “direct path loads” is disabled, the performance will be much worse than with the regular method of inserting data into Oracle (i.e. executing
INSERT
statements).Download the package Oracle Database Client (you do not need the installer of the full database). Download it from the URLs below:
For Oracle 19c or older:
If the Denodo server runs on Windows: Instant Client Downloads for Microsoft Windows (x64) 64-bit
If the Denodo server runs on Linux: Instant Client Downloads for Linux x86-64 (64-bit).
Install this package on the host where the Virtual DataPort server runs.
During the installation, select the Administrator installation type, which includes the program
sqlldr
. This is necessary to do bulk data loads.The default options of the installer are correct.
Edit the JDBC data source and do this:
In the tab Connection, verify that the connection URL is like this:
jdbc:oracle:thin:@<host>[:<port>]/<service>
The URLs below may cause the bulk loads to fail:
jdbc:oracle:thin:[<user>/<password>]@<host>[:<port>]:<SID> jdbc:oracle:thin:[<user>/<password>]@<TNSName>
Click the tab Read & Write. Then, select Use bulk data load APIs and in the Sqlldr executable location box, enter the path to the file
sqlldr
. This file is located where the “Oracle Database Client” was installed.