Hi all I noticed the following beaviour, if a command in a sql function fails the other statements fail like if the entire function where surrounded by a begin-commit block. While I find this behaviour nice, as far as I could find, this behaviour is not documented[1]. I wished to create a function that depends on such way of working but I cannot trust that sql functions will continue to work that way on the next postgresql versions.
So, is it documented somewhere?
I'm linking to a pastebin file that exemplifies the behaviour.
[1] -
http://www.postgresql.org/docs/9.1/static/sql-createfunction.html [2] -
http://pastebin.com/LWfFgH7L--
Marcelo Lacerda