So..are you implying that we fake an error to force the commit on a savepoint..and just ignore the error and move on..if that works..i guess it will work for me.
On Wed, Jul 15, 2009 at 5:49 PM, Scott Marlowe <scott.marl...@gmail.com>wrote: > So, did this not work either? > > > http://www.postgresql.org/docs/8.4/static/plpgsql-control-structures.html#PLPGSQL-ERROR-TRAPPING > > I had just figured that that hadn't worked, so you were looking for > something else. > > On Wed, Jul 15, 2009 at 4:32 PM, Anj Adu<fotogra...@gmail.com> wrote: > > No > > > > Savepoints allow a transaction to be committed upto a savepoint "if" > > somthing bad happens after savepoint..but I want an explicit commit even > if > > nothing bad happens later..i.e.I need to commit "when" I want to. (just > like > > Oracle PLSQL) > > > > On Wed, Jul 15, 2009 at 3:01 PM, Scott Marlowe <scott.marl...@gmail.com> > > wrote: > >> > >> On Wed, Jul 15, 2009 at 3:58 PM, Anj Adu<fotogra...@gmail.com> wrote: > >> > Is there a way to issue a commit within a function and continue > >> > processing? > >> > I know this is not possible in 8.1.. > >> > > >> > We have long running function calls...parts of which are independent > of > >> > each > >> > other. The independent functions are called within a "main" outer > >> > function. > >> > We would like to commit the data modified by each function as a single > >> > unit > >> > so that the data is freed up for other modules to act on. > >> > > >> > Are there any future postgres releases (we use 8.1.9) that possible > >> > address > >> > this. > >> > >> What you are looking for are savepoints, most likely. > > > > > > > > -- > When fascism comes to America, it will be intolerance sold as diversity. >
