I have an application using Oracle 10gR2 on linux. The requirement involves -
1. Reading large data files(upto 1 GB) and loading them into Oracle tables
2. Extracting large data(possibly millions of records) from Oracle tables and writing to data files
The input/output files could be in XML, CSV or some proprietary formats(involving non-standard end-of-record identifiers and field delimiters).
Considering the large I/O and database operations involved, I was considering 2 options:
1. Oracle java stored procedures
2. Pro*C
My understanding is that Pro*C should run faster and should therefore be preferred. At the same time, most applications nowadays appear to go for Oracle java stored procedures. Can you please highlight the pros and cons of these 2 approaches?
Which approach is better from the following perspectives -
1. Performance
2. Maintainability
3. Features/Capability
4. Portability (I assume OJSP wins here)