I've got a batch job that running more than 14hrs+ every day.
The Job designed such way that it creates 10 temporary SQL files (example file1.sql....file10.sql) using shell script
And each file will have 2000 query executions with unique i/p parameters to the script(qury.sql) something like below(@qury.sql 1 X 100).
The @qury.sql, which in turn spool the CSV data to a directory.
The file1.sql..file10.sql will start at a time and it's @qury.sql script will run serially for 2000 times.
So totally qury.sql has to execute 20,000 times serially. This is where most of the time being consumed
Is it possible to design to run @qury.sql in N parallel threads from file1.sql..file10.sql?
file1.sql
@qury.sql 1 X 100
@qury.sql 2 Y 300
@qury.sql 3 Z 400
@qury.sql 4 A 600
.
.
.
.
.
.
.
.
.
.
.
.
.
@qury.sql 2000 A 600
Please note batch job runs from client.
Oracle Version : 11.2.0.4.0