Turning off the shared cache seems to have solved the problem for most users, but one win32 user continues to report problems. I noticed that in 3.6.12 the default page size is automatically calculated on Windows. In my application I explicitly set the page size to 4096. Can having a page size that doesn't match the disk geometry cause problems? That is the only reason I can think of that the user continues to report problems that aren't reproducible with the same DB file here. To recap, the user gets errors like:
sqlalchemy.exc.IntegrityError: (IntegrityError) PRIMARY KEY must be unique u'update cards set priority = ? where id = ?' [[1, -9223199285979494924L], [1, -9221822696858457935L], [1, -9220362552298800344L], [1, -9218865005459903182L], [1, -9218053570259598995L], [1, -9217626953400592469L], [1, -9217257525142991358L], [2, -9217039826750418600L], [1, -9217011234538438799L], [1, -9216054651420921523L], [1, -9215471921529813571L], [3, -9215405945578177558L], [1, <strip a very long list> But if the user saves the deck and sends it to me, all the ids are unique. Cheers, Damien On Wed, Mar 25, 2009 at 9:38 AM, Damien Elmes <reso...@ichi2.net> wrote: > I can define the primary key column as not null if you think that will > help, but dumping the table reveals the ids are being assigned > sequential integers. > > On Tue, Mar 24, 2009 at 11:34 PM, Jim Wilcoxson <pri...@gmail.com> wrote: >> Not sure if it will make a difference, but in your trigger stuff you >> explicitly coded null for the primary key value. Have you tried >> changing that so that you don't specify the primary key field at all? >> I can't remember from the previous post, but I think it was (or should >> be) set up as autoincrement. >> >> I think SQLite allows using multiple nulls for the primary key, but >> according to their docs, it is non-standard and it says something >> about "this may change in the future". Maybe you are getting caught >> in the middle of a change that is going to occur across multiple >> revisions of SQLite. >> >> Jim >> >> >> On 3/24/09, Damien Elmes <reso...@ichi2.net> wrote: >>> Sorry, my application's files are called decks, and I unwittingly used >>> the wrong terminology. >>> >>> Any ideas about the problem? >>> >>> On Sat, Mar 21, 2009 at 2:27 AM, Griggs, Donald >>> <donald.gri...@allscripts.com> wrote: >>>>>>>> >>>>>>>>> However, when I ask the user to send me their deck, I find that: >>>>>>>>> >>>>>>>>> sqlite> pragma integrity_check; >>>>>>>>> integrity_check >>>>>>>>> --------------- >>>>>>>>> ok >>>>>>>>> sqlite> select id, count(id) from cards group by id having >>>>>>>>> count(id) >>>>>>>>>> 1; >>>>>>>>> sqlite> >>>>>>>>> >>>>>>>>> Any ideas? >>>> >>>> Obviously, that user is not playing with a full deck. ;-) >>>> >>>> >>>> _______________________________________________ >>>> sqlite-users mailing list >>>> sqlite-users@sqlite.org >>>> http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users >>>> >>> _______________________________________________ >>> sqlite-users mailing list >>> sqlite-users@sqlite.org >>> http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users >>> >> >> >> -- >> Software first. Software lasts! >> _______________________________________________ >> sqlite-users mailing list >> sqlite-users@sqlite.org >> http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users >> > _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
