On Tue, 04 Sep 2007 09:20:28 -0700, Gerry wrote: >Kees Nuyt wrote: >> On Tue, 04 Sep 2007 07:53:08 -0600, Scott wrote: >> >>> I have an application that is inserting a record every second. There >>> are thousands of periods from a few seconds to hours long where the data >>> in all the columns is identical thus causing hundreds of thousands of >>> duplicate rows. >>> >>> Is there a way to set up the INSERT statement to allow the insertion of >>> a new row of data, where one or more columns has data != to the previous >>> insertion, but ignore or skip the insertion of a duplicate row? >>> >>> Scott >> >> Perhaps: Specify all columns in the primary key, or define a >> separate table constraint UNIQUE (on,all,columns) ON CONFLICT >> IGNORE. >> The insert will fail. It depends on the conflict-clause what >> happens. Or you can use INSERT ... ON CONFLICT IGNORE. >> >> http://www.sqlite.org/lang_conflict.html >> >This may do what is needed, or it may not. > >This would skip an insertion that matched any previous data, even from >days earlier with many samples of different data in between. If the data >contents don't allow that, your scheme will work. > >Otherwise I think a TRIGGER could do what is needed. > >HTH, > >Gerry
You are right. Regards, -- ( Kees Nuyt ) c[_] ----------------------------------------------------------------------------- To unsubscribe, send email to [EMAIL PROTECTED] -----------------------------------------------------------------------------
