[Zeitgeist] [Bug 639737] Re: Improve insertion times

2011-02-10 Thread Seif Lotfy
** Changed in: zeitgeist Status: In Progress = Fix Committed ** Changed in: zeitgeist Status: Fix Committed = Fix Released -- You received this bug notification because you are a member of Zeitgeist Framework Team, which is subscribed to Zeitgeist Framework.

[Zeitgeist] [Bug 639737] Re: Improve insertion times

2011-01-19 Thread Seif Lotfy
** Changed in: zeitgeist Milestone: None = 0.8.0 -- You received this bug notification because you are a member of Zeitgeist Framework Team, which is subscribed to Zeitgeist Framework. https://bugs.launchpad.net/bugs/639737 Title: Improve insertion times Status in Zeitgeist Framework:

[Zeitgeist] [Bug 639737] Re: Improve insertion times

2010-11-16 Thread Seif Lotfy
OK I think the only way we can actually get the insertion time improved ATM is by using WAL for journaling. However it requires sqlite 3.7 and once its set to WAL sqlite 3.6 wont be able to access it. This breaks compatibility with older zeitgeist DBs: However I think this should be doable if we

[Zeitgeist] [Bug 639737] Re: Improve insertion times

2010-10-26 Thread Seif Lotfy
** Changed in: zeitgeist Milestone: 0.6 = 0.7 -- Improve insertion times https://bugs.launchpad.net/bugs/639737 You received this bug notification because you are a member of Zeitgeist Framework Team, which is subscribed to Zeitgeist Framework. Status in Zeitgeist Framework: In Progress

[Zeitgeist] [Bug 639737] Re: Improve insertion times

2010-10-19 Thread Seif Lotfy
OK I worked out the synchronous = NORMAL pragma for a week now and actually broke my netbook. What I learned is you have a better chance of messing up your partition than using synchronous = NORMAL. The deal breaker is the improvement is not noticeable unless we use journal = WAL However we are

[Zeitgeist] [Bug 639737] Re: Improve insertion times

2010-09-25 Thread Siegfried Gevatter
** Changed in: zeitgeist Milestone: None = 0.6 -- Improve insertion times https://bugs.launchpad.net/bugs/639737 You received this bug notification because you are a member of Zeitgeist Framework Team, which is subscribed to Zeitgeist Framework. Status in Zeitgeist Framework: In Progress

[Zeitgeist] [Bug 639737] Re: Improve insertion times

2010-09-18 Thread Seif Lotfy
** Changed in: zeitgeist Status: Triaged = In Progress -- Improve insertion times https://bugs.launchpad.net/bugs/639737 You received this bug notification because you are a member of Zeitgeist Framework Team, which is subscribed to Zeitgeist Framework. Status in Zeitgeist Framework: In

[Zeitgeist] [Bug 639737] Re: Improve insertion times

2010-09-17 Thread Seif Lotfy
Here a more detailed benchmarking: 1) NO CHANGES: ~ 0.12 s DEBUG:zeitgeist.engine:Inserted 1 events in 0.107999s DEBUG:zeitgeist.engine:Inserted 1 events in 0.155844s DEBUG:zeitgeist.engine:Inserted 1 events in 0.128242s DEBUG:zeitgeist.engine:Inserted 1 events in

[Zeitgeist] [Bug 639737] Re: Improve insertion times

2010-09-16 Thread Seif Lotfy
Here is a diff... https://code.edge.launchpad.net/~zeitgeist/zeitgeist/sqlite-insert-improvements/+merge/35701 Here I only set the journal_mode to off since we dont use rollback thus we can not be harmed AFAIK. It improved my average speed to 0.075 seconds which is better. I also benchmarked

[Zeitgeist] [Bug 639737] Re: Improve insertion times

2010-09-16 Thread Seif Lotfy
** Changed in: zeitgeist Assignee: (unassigned) = Seif Lotfy (seif) ** Changed in: zeitgeist Status: New = Triaged ** Changed in: zeitgeist Importance: Undecided = Wishlist ** Description changed: We insert pretty slowly with an average of 0.15 seconds for one event on my

[Zeitgeist] [Bug 639737] Re: Improve insertion times

2010-09-16 Thread Seif Lotfy
Markus, Table has exactly 5701 events and need 0.15 seconds per one event for insertion An empty table needs 0.1 seconds here. The time spent in python is too little to be mentioned. Sadly its the commit that takes most of the time. Inserting 10 events in my 5701 events database takes 1.0

[Zeitgeist] [Bug 639737] Re: Improve insertion times

2010-09-16 Thread Seif Lotfy
I think however our lower boundary is 0.05 seconds since we do commits after insert events is called, which is a disk write. Now we could avoid that by first returning the values to the user then commiting... It will save noticeable time. Also: I tried setting journal_mode to WAL Speed

[Zeitgeist] [Bug 639737] Re: Improve insertion times

2010-09-16 Thread Seif Lotfy
After reading more into the synchronous mode I found the mode NORMAL When synchronous is NORMAL (1), the SQLite database engine will still pause at the most critical moments, but less often than in FULL mode. There is a very small (though non-zero) chance that a power failure at just the wrong

Re: [Zeitgeist] [Bug 639737] Re: Improve insertion times

2010-09-16 Thread Seif Lotfy
What about journal_mode = WAL I see it being a good candidate too... On Thu, Sep 16, 2010 at 11:28 PM, Mikkel Kamstrup Erlandsen mikkel.kamst...@gmail.com wrote: Everyone be sure to read the details here http://www.sqlite.org/pragma.html so you can make an informed opinion. Here's what I