I have a sequential file on my work system (OS/390). This file contains
multiple record types which I want to load into multiple tables in a
PostgreSQL database. I have written a program which runs on the OS/390
system. This program reads the sequential file and reformats it. What I do
is reformat it so that a "copy <tablename> from stdin;" is generated
followed by the reformatted records for that table. This works and is
quite fast. However, I was wondering if anybody would know how much slower
it would be to use standard SQL INSERT statements would be. Basically, my
input file would look something like:
BEGIN;
INSERT .....
INSERT .....
COMMIT;
Again, I kinda like this approach because it is "standard" SQL and not
specific to PostgreSQL (not that I plan to use anything else,
personally). The reason I'd like a standard method is because there might
be some interest by other OS/390 users in this code for other databases
which they might have.
Just curious,
John McKown