Using SQL*Loader and named pipes

[djlocke]

Hello!

Has anyone ever tried using orabulk stage in conjunction with a named pipe in UNIX as a target? I want to avoid first writing the data file and in a second step having to execute sqlldr. With a named pipe I assume I could do it in one step by letting DataStage write into a named pipe and sqlldr read out of it.

Any comments?

Juergen

 

[Guest_imported]

I just went through this during a DataStage XE 5.2 evaluation. You can use pipes with the OraOCI8 Bulk Loader provided you're running the DataStage server on the same machine (and same domain if the unix box has multiple domains) as the Oracle database. If you have this configuration you can bypass SQLLoader and DataStage's bulk loader will write directly to the Oracle database at a low level.

As it doesn't fit our environment, we won't be using this feature.

- rob