Hey Thomas, > There are other ways to improve performance. Did you already try out > some other options? How much do you need to improve the speed, do you > have a specific performance goal?
I also am currently testing oracle's berkeley db java edition, firebird, and sybase's sql anywhere. At the moment, it seems like H2 and berkeley db are close in performance in terms of inserts on a clean database. I have not done thorough tests, but it seems that berkeley db does better for my application and my data I think because of the clustered index and the always increasing nature of the timestamp records probably only need to be appended and indexing isn't too difficult. The only option I've tried to improve speed was the partitioning of the data domain, and that hasn't gone well as you can tell from my other post. Like I said, H2's speed is great, but once I hit a certain size the performance really really takes a hit. I only need to improve the speed when my database is large. So it's more of a scalability issue. My performance goal for H2 is to beat berkeley. I can get 20 writes/ ms with transactions that don't sync to disk, or I can get 55 writes/ ms with deferred non transactional writes. Now from what I've seen H2 easily beats those numbers, but not when the database grows. I can tell you more if you'd like. At the moment I'm doing more thorough benchmarks that I've created. Thanks for the feedback. Julian On Oct 23, 4:47 pm, "Thomas Mueller" <[EMAIL PROTECTED]> wrote: > Hi, > > > From what I understand, a non-clustered index will store a key-value > > mapping outside of the database where the keys are ordered and the > > values are locations of records. A clustered index will actually > > store records in sorted order within the database itself, or so I > > think. > > There is a feature like that on the roadmap, "Index organized tables > CREATE TABLE...(...) ORGANIZATION INDEX (store in data file)". > > Like you wrote, it should improve performance if you access the table > using this index (the 'clustered index'). Unfortunately to implement > it, a file format change is required. For some cases it will actually > decrease performance however, even when accessing the table using this > index (for example COUNT(*), MIN(column), MAX(column) with a where > clause on the column), so I'm not sure how much it will help in your > case. > > > Hopefully I'm asking a reasonable question. > > Sure! > > There are other ways to improve performance. Did you already try out > some other options? How much do you need to improve the speed, do you > have a specific performance goal? > > Regards, > Thomas --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "H2 Database" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/h2-database?hl=en -~----------~----~----~----~------~----~------~--~---
