Re: [sqlite] Very, very slow commits

Hi Rob, Answers are in the text below On Mon, Jul 30, 2018, at 13:45, Rob Willett wrote: > Droedel, > > We don't think there are significant read access. The database is a > single database on a single thread on a single process. The only access > to it is a Perl script th

Re: [sqlite] Very, very slow commits

uldn't find a pragma for this. Oh, and do both databases have the same page size, preferably 4K ? Regards, Droedel On Mon, Jul 30, 2018, at 12:51, Rob Willett wrote: > Droedel, > > Thanks for the comprehensive reply. We have actually done all of this. > > The system has

Re: [sqlite] Very, very slow commits

isn't running. Just my 2 cents. Kind regards, Droedel On Sun, Jul 29, 2018, at 10:14, Rob Willett wrote: > Hi, > > Background > > We've been refactoring our database to reduce the size of it. Through > some simple logic we've managed to pull out 99% of the dat

[sqlite] Slow real world performance - Any suggestions please (warning long)

Rob Willett writes: [snip] > The headline figures are we have gone from 213 secs to process 20 files down to 90 secs to process 20 files. We > are running approx 2.5x faster. To get this improvement the biggest change was simply adding COLLATE > NOCASE to the table schema. This saved around 120-1