Re: Could postgres12 support millions of sequences? (like 10 million)

On Mar 19, 2020, at 7:35 PM, pabloa98 <pabloa98@gmail.com> wrote:

On Thu, Mar 19, 2020 at 6:16 PM Rob Sargent <robjsargent@gmail.com> wrote:

On Mar 19, 2020, at 6:45 PM, pabloa98 <pabloa98@gmail.com> wrote:

On Thu, Mar 19, 2020 at 5:13 PM Adrian Klaver <adrian.klaver@aklaver.com> wrote:

On 3/19/20 3:32 PM, pabloa98 wrote:
> 
> 
> On Thu, Mar 19, 2020 at 3:17 PM Rob Sargent <robjsargent@gmail.com 
> <mailto:robjsargent@gmail.com>> wrote:I have a table called "pair":

CREATE TABLE pair(

group INT NOT NULL,

element INT NOT NULL,

CONSTRAINT PRIMARY KEY (group, element)

);

I must add entries to the table "event". This table event will have a code that will be generated using a sequence of that is a function of s(group,element). 

CREATE TABLE event(

group INT NOT NULL,

element INT NOT NULL,

code INT NOT NULL,

CONSTRAINT PRIMARY KEY(code, element, group)

);

Unless event table is searched by code more than group, you probably want to maintain the key order from the pair table's primary key.

OK. I will do that. 

If gaps are ok do you still near multiple sequences?

I need to start each "code" value from 1 in each (group, element) pair. This is because we only have 99999999 codes for each pair. We do not want to waste numbers. If there is a gap is OK but no a gap of millions of numbers.