- Архив списков рассылки pgsql-interfaces

Hi there.

I'm running Postgresql 7.3.2 on Mandrake 9.1.
I'm using ecpg (version 2.10.0) to compile the embedded sql code in my C
programs.

The problem experienced has to do with array indexing in SQL statements.
I browsed your test files in the CVS repository and found one
(http://developer.postgresql.org/cvsweb.cgi/pgsql-server/src/interfaces/
ecpg/test/test1.pgc?rev=1.31.2.1&content-type=text/x-cvsweb-markup) that
contains the code I would like to be able to use.

When I precompile the abovementioned file with ecpg, I get the following
error test1.pgc:130: parse error at or near "[".

It is the same problem I'm experiencing with my program. I tested my
program on Postgresql 7.3.4 as well, but it also did not work (the ecpg
version was the same as the 7.3.2's).

What do you suggest I do to get this working? Grab latest code from CVS?

My second question is with regards to Oracle style bulk inserts/updates.
I've played around with the "COPY" command and it seems to have good
performance (even though I have to explicitly sprintf all values in a
string first). I'm currently working on a port of existing software that
stores huge amounts of (telecoms) data, so performance is critical. Is
there any special way of handling a lot of updates? I can understand
that updates might need to be done one at a time, but it would be nice
if I can cut out the overhead of calling an update sql statement
repeatedly - I'm thinking in the line of the Oracle bulk updates where
you give the sql statement and host arrays containing the fieldvalues
for the different updates.

Any hints/suggestions would be appreciated.

Regards, Cobus Carstens