Re: [sqlite] sqlite abnormal IO writing

2013-04-22 Thread Stephen Chrzanowski
Just for "fun" try turning synchronous = OFF and see what kind of speed boost you get. On Sun, Apr 21, 2013 at 9:41 PM, 刘运杰 wrote: > > > My database connection option: > > > PRAGMA synchronous = NORMAL; > PRAGMA journal_mode = OFF; > PRAGMA auto_vacuum = 0; > > > In windows process monitor ,it

Re: [sqlite] sqlite abnormal IO writing

2013-04-21 Thread 刘运杰
My database connection option: PRAGMA synchronous = NORMAL; PRAGMA journal_mode = OFF; PRAGMA auto_vacuum = 0; In windows process monitor ,it show 2.5M writing ,but the database file increase only 10k or so. So those IO writing is very abnormal IMO. At 2013-04-22 03:24:43,"Stephen Chr

Re: [sqlite] sqlite abnormal IO writing

2013-04-21 Thread 刘运杰
Here is the schema of table: CREATE TABLE statdata ( typeINT, seq INT, uid VARCHAR( 36 ), counter INT, PRIMARY KEY ( type, seq ASC ) ); before my modify: 500 rows at maximum, query:insert or replace (...) in batch mode take 10-20

Re: [sqlite] sqlite abnormal IO writing

2013-04-21 Thread Klaas V
>On Sun, 21 Apr 2013 11:15:23 +0800 (CST), 刘运杰 >wrote: >Ok,I do not make my means clearly. I mean 60 seconds after my >program started,not token 60 seconds to load database file. ... >Now,I modify the implement of batch query, it take about one >second and 70k IO writing.So there are abnormal some

Re: [sqlite] sqlite abnormal IO writing

2013-04-21 Thread Stephen Chrzanowski
You know, this kind of "smells" like a full sync, non transactioned, single inserts problem. Basically, you open the connection to the database, start throwing inserts at it without a transaction around each insert. To make matters worse (For time anyways) you've probably got full sync on, and yo

Re: [sqlite] sqlite abnormal IO writing

2013-04-21 Thread Paolo Bolzoni
FAQ 19 applies also to SAVEPOINT/RELEASE, not only to BEGIN/COMMIT, right? On Sun, Apr 21, 2013 at 9:35 AM, Kees Nuyt wrote: > On Sun, 21 Apr 2013 11:15:23 +0800 (CST), ?? > wrote: >> >> Ok,I do not make my means clearly. I mean 60 seconds after my >> program started,not token 60 seconds to

Re: [sqlite] sqlite abnormal IO writing

2013-04-21 Thread ibrahim
On 21.04.2013 05:15, 刘运杰 wrote: Ok,I do not make my means clearly. I mean 60 seconds after my program started,not token 60 seconds to load database file. Now, I got the reason of sqlite abnormal IO writing,it about batch insert. Here is my usecase: One table about 4 column and 500 row,the conte

Re: [sqlite] sqlite abnormal IO writing

2013-04-21 Thread Clemens Ladisch
刘运杰 wrote: >One table about 4 column and 500 row,the content of every row no exceed >100 byte, every time I update the whole table using batch query. > >It should take about one second and 100k IO writing on estimate,BUT it >sustained about 20 second and wrote about 2.5M actually. Please show that

Re: [sqlite] sqlite abnormal IO writing

2013-04-21 Thread Kees Nuyt
On Sun, 21 Apr 2013 11:15:23 +0800 (CST), ?? wrote: > > Ok,I do not make my means clearly. I mean 60 seconds after my > program started,not token 60 seconds to load database file. > > Now, I got the reason of sqlite abnormal IO writing,it about > batch insert. Here is my usecase: One table ab

Re: [sqlite] sqlite abnormal IO writing

2013-04-20 Thread 刘运杰
Ok,I do not make my means clearly. I mean 60 seconds after my program started,not token 60 seconds to load database file. Now, I got the reason of sqlite abnormal IO writing,it about batch insert. Here is my usecase: One table about 4 column and 500 row,the content of every row no exceed 100 byt

Re: [sqlite] sqlite abnormal IO writing

2013-04-20 Thread Stephen Chrzanowski
Attachments cannot be put into this mailing list. Please either upload to a service like PasteBin or put a copy in your PUBLIC directory in DropBox and provide us with a link. Try using a different application, like the Command Line Interface (CLI) and open it through there and see if you get a h