Обсуждение: int4/int8 problem with jdbc was Re: "Fix" for INT8 literals being parsed as INT4 disqualifying
hi all, i have some problems with my java app and int8 pks. i browsed the mail-archives and found the following mail thread <bold><fontfamily><param>Times</param><bigger><bigger>Re: "Fix" for INT8 literals being parsed as INT4 disqualifying</bigger></bigger></fontfamily></bold> the last statement was: <fontfamily><param>Courier</param><x-tad-bigger>That patch will be applied soon. But it will *not* be in the 7.4 stable branch. It will be fixed in head of tree. Since it will break existing code I didn't want to add it to 7.4 in the middle of beta, or now that 7.4 is RC1. </x-tad-bigger></fontfamily> full text of the mail i'm referring to: http://archives.postgresql.org/pgsql-jdbc/2003-11/msg00057.php has this been already adressed in the HEAD for the jdbc-driver? because i loaded it from gborg via cvs and the problem still persists. http://gborg.postgresql.org/project/pgjdbc/projdisplay.php tia, jan hi all, i have some problems with my java app and int8 pks. i browsed the mail-archives and found the following mail thread Re: "Fix" for INT8 literals being parsed as INT4 disqualifying the last statement was: That patch will be applied soon. But it will *not* be in the 7.4 stable branch. It will be fixed in head of tree. Since it will break existing code I didn't want to add it to 7.4 in the middle of beta, or now that 7.4 is RC1. full text of the mail i'm referring to: http://archives.postgresql.org/pgsql-jdbc/2003-11/msg00057.php has this been already adressed in the HEAD for the jdbc-driver? because i loaded it from gborg via cvs and the problem still persists. http://gborg.postgresql.org/project/pgjdbc/projdisplay.php tia, jan
Jan Heise wrote: > hi all, > > i have some problems with my java app and int8 pks. i browsed the > mail-archives and found the > following mail thread *Re: "Fix" for INT8 literals being parsed as INT4 > disqualifying* > the last statement was: > > That patch will be applied soon. But it will *not* be in the 7.4 stable > branch. It will be fixed in head of tree. Since it will break existing > code I didn't want to add it to 7.4 in the middle of beta, or now that > 7.4 is RC1. > > full text of the mail i'm referring to: > > http://archives.postgresql.org/pgsql-jdbc/2003-11/msg00057.php > > has this been already adressed in the HEAD for the jdbc-driver? because > i loaded it from > gborg via cvs and the problem still persists. I don't believe the patch in question ever got applied. The 7.5 server doesn't suffer from the same "type mismatch prevents use of an index" problem, I believe. Or you could add an explicit ::int8 cast to your query. My v3 protocol changes should fix this against a 7.4 server, but they're some distance away from making it into the official driver at the moment. -O
oliver, thanks for your reply. the db in question is a 7.4 site in production. upgrading to another version is not really what i'd like to do. the explicit cast is problematic too, i'm using hibernate as my orm. unless i missed something in the docs, i can not add the cast to the config - but this is sth. for the hibernate forums. the problem is a special case where i have to load objects from the table in question by the pk (details would be quite long, but i can not change the table-structure at this time). table size is approx. 120 million rows and approx 2 million rows have a pk < INT4_MAX. so i think my only option is dropping the foreign keys, making the pks and all references bigger than INT4_MAX and recreate the fk. jan Am 02.06.2004 um 13:16 schrieb Oliver Jowett: > Jan Heise wrote: >> hi all, >> i have some problems with my java app and int8 pks. i browsed the >> mail-archives and found the >> following mail thread *Re: "Fix" for INT8 literals being parsed as >> INT4 disqualifying* >> the last statement was: >> That patch will be applied soon. But it will *not* be in the 7.4 >> stable branch. It will be fixed in head of tree. Since it will break >> existing code I didn't want to add it to 7.4 in the middle of beta, >> or now that 7.4 is RC1. >> full text of the mail i'm referring to: >> http://archives.postgresql.org/pgsql-jdbc/2003-11/msg00057.php >> has this been already adressed in the HEAD for the jdbc-driver? >> because i loaded it from >> gborg via cvs and the problem still persists. > > I don't believe the patch in question ever got applied. > > The 7.5 server doesn't suffer from the same "type mismatch prevents > use of an index" problem, I believe. Or you could add an explicit > ::int8 cast to your query. > > My v3 protocol changes should fix this against a 7.4 server, but > they're some distance away from making it into the official driver at > the moment. > > -O > > ---------------------------(end of > broadcast)--------------------------- > TIP 6: Have you searched our list archives? > > http://archives.postgresql.org
On 02/06/2004 12:47 Jan Heise wrote: > oliver, thanks for your reply. > > the db in question is a 7.4 site in production. upgrading to another > version is not really what i'd like > to do. the explicit cast is problematic too, i'm using hibernate as my > orm. unless i missed something > in the docs, i can not add the cast to the config - but this is sth. for > the hibernate forums. > > the problem is a special case where i have to load objects from the > table in question by the pk > (details would be quite long, but i can not change the table-structure > at this time). table size is > approx. 120 million rows and approx 2 million rows have a pk < INT4_MAX. > so i think my only > option is dropping the foreign keys, making the pks and all references > bigger than INT4_MAX > and recreate the fk. > > jan There's an unofficial patch: http://archives.postgresql.org/pgsql-jdbc/2003-09/msg00143.php which might solve your problems for now. IIRC, there's a server-side patch been submitted which will solve the int4/int8 issue and may be 7.5. -- Paul Thomas +------------------------------+---------------------------------------------+ | Thomas Micro Systems Limited | Software Solutions for Business | | Computer Consultants | http://www.thomas-micro-systems-ltd.co.uk | +------------------------------+---------------------------------------------+
Thanks for your help, for this special case I patched the JDBC-driver to wrap setInt and setLong to use setString instead. Jan Am 02.06.2004 um 15:27 schrieb Paul Thomas: > > On 02/06/2004 12:47 Jan Heise wrote: >> oliver, thanks for your reply. >> the db in question is a 7.4 site in production. upgrading to another >> version is not really what i'd like >> to do. the explicit cast is problematic too, i'm using hibernate as >> my orm. unless i missed something >> in the docs, i can not add the cast to the config - but this is sth. >> for the hibernate forums. >> the problem is a special case where i have to load objects from the >> table in question by the pk >> (details would be quite long, but i can not change the >> table-structure at this time). table size is >> approx. 120 million rows and approx 2 million rows have a pk < >> INT4_MAX. so i think my only >> option is dropping the foreign keys, making the pks and all >> references bigger than INT4_MAX >> and recreate the fk. >> jan > > There's an unofficial patch: > > http://archives.postgresql.org/pgsql-jdbc/2003-09/msg00143.php > > which might solve your problems for now. IIRC, there's a server-side > patch been submitted which will solve the int4/int8 issue and may be > 7.5. > > -- > Paul Thomas > +------------------------------ > +---------------------------------------------+ > | Thomas Micro Systems Limited | Software Solutions for Business > | > | Computer Consultants | > http://www.thomas-micro-systems-ltd.co.uk | > +------------------------------ > +---------------------------------------------+ > > ---------------------------(end of > broadcast)--------------------------- > TIP 5: Have you checked our extensive FAQ? > > http://www.postgresql.org/docs/faqs/FAQ.html