Ok... you are using SQL unload/load to do this...
This command is very simple to use, but is not have a good performance..
You have a lot of options, you can try tune your load/unload SQL, use dbload, use HPL.
My suggestion is HPL ( High Performance Load).
I never work with a version before the 7.31. I Don't know how HPL works exactly in version 7.2
To get a gain in this process if possible use more devices (physical disks).
The HPL commands to use are;
- ipload : Here you create a job to load/unload your table. This is a GUI application , you need a XDMCP client (*) or your must have a graphical enviroment on your server.
- onpload : Execute the job.
(*) I know 2 ways to use a "XDMCP client" :
1) download a XDMCP client for windows on the internet, i recommend XDEEP,
2) Use a linux station (can be a live linux, bootable), start the KDE/GNOME and active connections with "xhosts +"
Before execute ipload, on the same session you must set the enviroment DISPLAY=<ip>:0 , eg: DISPLAY=172.30.252.145:0
When you execute the ipload the application will popup in you client XDMCP
What you need to do:
1) check how much physical devices you have available to use, if are mounted on you file system.
2) start your XDMCP client
3) Open a new session in your server and set DISPLAY variable to IP of XDMCP client
4) execute the ipload
5) On the ipload, create the new "device", you will create a file path to each physical disk. When export occur , the HPL split your data into this files with parallel unload.
6) create the job. I don't know if have the option "no convertion" on the HPL of IDS 7.2 , if have, set, it's faster.
7) to unload: onpload -j <job_name> -fu
8) to load: onpload -j <job_name> -fl
The format exported by HPL (if you _not_ set the option "no conversion" ) is compatible with SQL load/unload.