On Sat, Jan 07, 2006 at 02:16:46PM -0500, [EMAIL PROTECTED] wrote:
> Isaac Raway <[EMAIL PROTECTED]> wrote:
> > [I]nstead of dropping the row existing, [the REPLACE algorithm should]
> > simply update the provided fields in place, leaving the unmodified fields
> > as is. I'd call this behavior OR
Hello!
I think, there is another way for high-concurrency SQLite-based systems.
**It is not entirely universal**, but I hope it may be used for
high-traffic web sites and similar kind of systems, where each individual
transaction (such as a page retrieval or a form submission) is very
short and
Alexander,
I like your general concept. Some years ago I implemented a somewhat
similar strategy in a product of ours which was SQLite-like in that it
linked into each application process and managed B-Trees. By
identifying read-only transactions and handling them in a simple manner
It will be always get 'database table is locked' message when trying to
delete
a record after select.
I'v tested on WinXP and Ubuntu 5.10, but got the same error.
SQLite version v3.2.8
please see the following code
#include
#include
#include
#define DATFILE "test.dat"
int main(int argc,
Justin Wu wrote:
It will be always get 'database table is locked' message when trying
to delete
a record after select.
You are not deleting after select - you are deleting while still
performing select. You can't do a write operation in the middle of a
read. You have to select, remember the
5 matches
Mail list logo