Re: cdr_serial issue

Thanks. I got this working and I understand now why they did it this way. You don't want every table to automatically burn serials just because you have the CDR_SERIAL id turned on. I wish you could specify serial id parameters like you can specify sequences. I can't be the only person who wanted to generate serials in a specific way. I mean they included it in the sequence specification so the standard committee thought this sort of thing was at least useful enough to make it part of the standard. I know just because it is part of the standard doesn't automatically make it a good thing. On Nov 9, 12:38 pm, "TBP (The Big Potato)" <T...@NotHere.Co.Uk> wrote: > bozon wrote: > > On Nov 9, 8:40 am, bozon <cur...@crowson1.com> wrote: > > >>On Nov 9, 8:13 am, bozon <cur...@crowson1.com> wrote: > > >>>On Nov 9, 6:54 am, "TBP (The Big Potato)" <T...@NotHere.Co.Uk> wrote: > > >>>>bozon wrote: > > >>>>>I made a mistake about cdr_serial I assumed that it would work > >>>>>regardless of whether ER was on or not. This doesn't seem to be the > >>>>>case. I need to spread out my serial IDs like CDR_SERIAL does. > > >>>>>Is there a way to get the functionality of CDR_SERIAL without actually > >>>>>replicating? > > Well, this is a hack, but it does work ... create a replicate with one participant : > > create database jj_one_repl in dbspace2 with log; > database jj_one_repl; > create table tab1 (col1 serial primary key, col2 char(25)) with crcols lock mode row;> EOF > cdr define repl -i -A -R -C timestamp -S trans jj_one_repl \\ > "jj_one_repl@nan_1000f_1_cdr:informix.tab1" "select * from tab1" > cdr start repl jj_one_repl > > and ensure you have CDR_SERIAL set to something sensible (2,0 for site 1 and 2,1 for site 2 for example if just 2 sites) > > Then you get the CDR_SERIAL behaviour (and ER doing a bit of work to realise it has nothing to do).