Re: [sqlite] linear behavior on INSERT (was Re: unexpected large journal file)

On 6/22/10, Eric Smith wrote: > Jim Wilcoxson wrote: ... >> Did you see my earlier note about combining your two integers into the >> primary key? This will also give you constant insert times, if you >> insert items in the order: ... > Thanks also for the tip on insertion

Re: [sqlite] linear behavior on INSERT (was Re: unexpected large journal file)

On 23 Jun 2010, at 1:18am, Eric Smith wrote: > *sigh* kill me. Sorry for wasting your time there. :/ We've all done it. Simon. ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users

Re: [sqlite] linear behavior on INSERT (was Re: unexpected large journal file)

Scott Hess wrote: > You should reduce your demonstration case to something you'd be > willing to post the code for. Probably using synthetic data > (preferably generated data). There's something missing in the thread > right now, and it's unlikely to be exposed by random shots in the >

Re: [sqlite] linear behavior on INSERT (was Re: unexpected large journal file)

Scott Hess wrote: > You should reduce your demonstration case to something you'd be > willing to post the code for. Probably using synthetic data > (preferably generated data). There's something missing in the thread > right now, and it's unlikely to be exposed by random shots in the >

Re: [sqlite] linear behavior on INSERT (was Re: unexpected large journal file)

Eric, You should reduce your demonstration case to something you'd be willing to post the code for. Probably using synthetic data (preferably generated data). There's something missing in the thread right now, and it's unlikely to be exposed by random shots in the dark. -scott On Tue, Jun

Re: [sqlite] linear behavior on INSERT (was Re: unexpected large journal file)

Jay A. Kreibich wrote: > What OS/filesystem are you using? > > SQL inserts should be near-constant, assuming the table does not > have an INTEGER PRIMARY KEY with explicit values. The table's root > B-Tree needs to re-balance every now and then, but if the inserts are > in-order (which

Re: [sqlite] linear behavior on INSERT (was Re: unexpected large journal file)

Uggg On Tue, Jun 22, 2010 at 05:12:38PM -0500, Jay A. Kreibich scratched on the wall: > On Tue, Jun 22, 2010 at 04:16:42PM -0400, Eric Smith scratched on the wall: > > Jim Wilcoxson wrote: > > > > > Insert times should be constant for the 2nd case: no primary key, no > > > indexes; ie,

Re: [sqlite] linear behavior on INSERT (was Re: unexpected large journal file)

On Tue, Jun 22, 2010 at 04:16:42PM -0400, Eric Smith scratched on the wall: > Jim Wilcoxson wrote: > > > Insert times should be constant for the 2nd case: no primary key, no > > indexes; ie, it doesn't matter how many records are already in the > > database. I confirmed this with SQLite

Re: [sqlite] linear behavior on INSERT (was Re: unexpected large journal file)

Richard Hipp wrote: > When there are no indices, SQLite can put the rows into the database in > any order it wants, and it chooses to put them in ascending order. Thus, > each insert becomes a constant-time append. (Approximately - the truth is > a little more complicated, but by waving our

Re: [sqlite] linear behavior on INSERT (was Re: unexpected large journal file)

On Tue, Jun 22, 2010 at 3:08 PM, Eric Smith wrote: > I have confirmed that INSERT times are roughly logarithmic in > the number of existing records after creating my specific user > indices. > > But INSERT times appeared to be *linear* in the number of existing > records

Re: [sqlite] linear behavior on INSERT (was Re: unexpected large journal file)

Jim Wilcoxson wrote: > Insert times should be constant for the 2nd case: no primary key, no > indexes; ie, it doesn't matter how many records are already in the > database. I confirmed this with SQLite 3.6.18. Definitely not constant. Looks linear to me -- you saw the plot, you can decide

Re: [sqlite] linear behavior on INSERT (was Re: unexpected large journal file)

On 6/22/10, Eric Smith wrote: > I have confirmed that INSERT times are roughly logarithmic in > the number of existing records after creating my specific user > indices. > > But INSERT times appeared to be *linear* in the number of existing > records before I had created any

[sqlite] linear behavior on INSERT (was Re: unexpected large journal file)

I have confirmed that INSERT times are roughly logarithmic in the number of existing records after creating my specific user indices. But INSERT times appeared to be *linear* in the number of existing records before I had created any user indices (and with no primary keys or unique indices