Re: [sqlite] 'insert or ignore' vs self join?

2008-05-28 Thread Petite Abeille
Hello, On May 27, 2008, at 9:07 PM, Stephen Oberholtzer wrote: > Well, the first thing you should bring away from this experience is > that the > number of VM instructions isn't really an indicator of how efficient > the > query is :) Good point :) > Now, I'm not sure exactly why one is

Re: [sqlite] 'insert or ignore' vs self join?

2008-05-27 Thread Stephen Oberholtzer
On Tue, May 27, 2008 at 2:41 PM, Petite Abeille <[EMAIL PROTECTED]> wrote: > Hello, > > % sqlite3 -version > 3.5.9 > > I'm trying to figure out a frugal way to handle a unique key > constrain... > > I tried using both 'insert or ignore' and a self join. The self join > seems to be noticeably

[sqlite] 'insert or ignore' vs self join?

2008-05-27 Thread Petite Abeille
Hello, % sqlite3 -version 3.5.9 I'm trying to figure out a frugal way to handle a unique key constrain... I tried using both 'insert or ignore' and a self join. The self join seems to be noticeably faster even though 'insert or ignore' would empirically appear to be the better deal