up. I don't believe it's a locking
issue since my timer doesn't start until the query is successfully executed.
Any ideas? Would the occasional write operation in the midst of these reads
cause any permanent slow down to the read time? Thanks.
Mike Borland
-Original Me
oretically call sqlite3_step() multiple times just to iterate rows?
I happen to be binding the data in each row, but I thought that is
optional.
Mike Borland
-Original Message-
From: Simon Slavin [mailto:slav...@hearsay.demon.co.uk]
Sent: Friday, June 12, 2009 6:00 PM
To: General Discussion of
lite3_step() takes less than a millisecond to run. After 0-50
iterations, it's taking anywhere from 10-100ms.
Does anybody have any insight into what's happening behind the scenes
with this function to help me track down the cause? I apprecia
have details on
when and how the cache is flushed?
Thanks,
Mike Borland
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
to
implement a fifo queue on my side to make sure that the writers obtained
the lock in the order requested. Otherwise I saw cases where one thread
kept getting the lock because it would finish its work and request the
lock again before the other waiting threads completed their busy timeout
idual tables. Also, the documentation seemed to indicate that it
had some relationship to the shared cache mode. Would enabling the
read_uncommited pragma ever result in invalid data being read from a
table which is being written to?
Thank for any help!
Mike Borland
CygNet Software, I
6 matches
Mail list logo
