Thread: RI violation msg suggestion
Minor usability/debuggability suggestion... RI violation error messages in 7.0.0 do not appear to identify the offending value. Example: ERROR: fk_employee_currency referential integrity violation - key referenced from employee not found in currency Easier to debug would be: ERROR: fk_employee_currency referential integrity violation - key '27' referenced from employee not found in currency Regards, Ed Loehr
Greetings! 'copy from stdin' on 7.0.2 appears to simply hang if more than ~ 1000 records follow in one shot. I couldn't see this behavior documented anywhere. Is this a bug? Take care, -- Camm Maguire camm@enhanced.com ========================================================================== "The earth is but one country, and mankind its citizens." -- Baha'u'llah
Camm Maguire <camm@enhanced.com> writes: > Greetings! 'copy from stdin' on 7.0.2 appears to simply hang if more > than ~ 1000 records follow in one shot. I couldn't see this behavior > documented anywhere. Is this a bug? I've never heard of any such behavior ... and you can be sure that we'd have heard about this, since any moderately large pg_dump file would trigger such a bug. You must have something else going on. Details please? regards, tom lane
Greetings, and thank you for your reply. OK, I have 4 tables, and a view on a merge of the 4. I have a trigger on insert into table 3, and a trigger on insert into the view, which basically just takes the input data, does a few selects on the tables, and inserts the appropriate portions of the data into each table as necessary. When I copy up to ~ 1000 lines of a file into this view, everything goes fine. More than that, and after a while, cpu activity stops, disk activity stops, and the job hangs indefinitely. Control-C gives the later message "You have to get out of copy state yourself". I can provide the schema if needed. Thanks for your help! Tom Lane <tgl@sss.pgh.pa.us> writes: > Camm Maguire <camm@enhanced.com> writes: > > Greetings! 'copy from stdin' on 7.0.2 appears to simply hang if more > > than ~ 1000 records follow in one shot. I couldn't see this behavior > > documented anywhere. Is this a bug? > > I've never heard of any such behavior ... and you can be sure that we'd > have heard about this, since any moderately large pg_dump file would > trigger such a bug. You must have something else going on. Details > please? > > regards, tom lane > > -- Camm Maguire camm@enhanced.com ========================================================================== "The earth is but one country, and mankind its citizens." -- Baha'u'llah
Camm Maguire <camm@enhanced.com> writes: > OK, I have 4 tables, and a view on a merge of the 4. I have a trigger > on insert into table 3, and a trigger on insert into the view, which > basically just takes the input data, does a few selects on the tables, > and inserts the appropriate portions of the data into each table as > necessary. > When I copy up to ~ 1000 lines of a file into this view, everything > goes fine. I'm a bit startled that COPY to a view works at all ;-). But it does look like copy honors triggers, so in principle the above ought to work. I'll need to see the complete details of the trigger and all the referenced table declarations --- probably don't need the data, though, unless the trigger needs to have nonempty tables to start with. regards, tom lane