What's wrong with the following (pseudo-code)?
begin immediate
x = result of (select max(j) from mytab;);
++x;
for (n=0; n<num; n++) // num may be greater than or equal to zero
{
insert into mytab (j) values (x);
}
commit
Just don't forget to create index on j.
Pavel
On Thu, Oct 27, 2011 at 3:12 PM, Tim Streater <[email protected]> wrote:
> Is there a way to get a unique id from a database without actually creating a
> new row? If I have:
>
> create table mytab (i integer primary key, j);
>
> then I could do (pseudo-code):
>
> insert into mytab (i) values (null);
> x = result of (select last_insert_rowid() from mytab;);
>
> for (n=0; n<num; n++) // num may be greater than or equal to zero
> {
> insert into mytab (j) values (x);
> }
>
> delete from mytab where i=x;
>
>
> This would give me my desired result of having a set of rows with the same
> j-value (other columns differ, of course). The insert in the for-loop is
> actually done elsewhere so it doing as above simplifies the logic. As num may
> be zero I may not in fact need the unique value. What I'd like would be a way
> of asking table mytab for a unique value that may or may not get used, but
> without the need to thereby create a row that I later have to delete.
>
> --
> Cheers -- Tim
>
> _______________________________________________
> sqlite-users mailing list
> [email protected]
> http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
>
>
_______________________________________________
sqlite-users mailing list
[email protected]
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
