We were having some trouble doing updates to our database,
a lot of our database sort of works like this:
dbfunc(data) somedatatype *data;
{somedatatype *existing_row;
existing_row = exists_in_table(data);
if (existing_row != NULL) { update_table(existing_row, count = count + data->count)} else
insert_into_table(data);
}
Is there anything built into postgresql to accomplish this without
the "double" work that goes on here?
something like: update_row_but_insert_if_it_doesn't_exist(data, update = 'count = count + data->count');
Meaning, if a row matching the 'new' data exists, update it, otherwise
store our new data as a new record?
It seems like the database has to do an awful amount of extra work
for our application because we haven't figured out how to do this
effeciently.
Any pointers?
thanks,
--
-Alfred Perlstein - [bright@wintelcom.net|alfred@freebsd.org]
