Strange but it seemed that it was related to multi-threading (although I am
not sharing DB pointers).
I had one thread writing and another one reading and the Reader thread
started reading before I could commit the transaction in the first thread.
(Though I thought SQLite would have a Pending Lock on the same). After the
long read was done, when the writer tried to commit it returned this error.
I put in my own locking mechanism and havent been able to reproduce it
since. Have I taken the right approach? I'd rather understand and fix the
problem sooner than later.
Also is there any way to view the Temp Table that I create programatically
in the Command Line?
Thanks!
On 10/25/05, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
>
> R S <[EMAIL PROTECTED]> wrote:
> > with Return Value 13.
> > I checked my partition space and its usage is just 2% (Platform is Linux
> > using SQLite 3.2.2).
> > Also I am using Temp tables and periodically move data into my Main
> Table. I
> > wondered if my Temp Table is full because strace on my process gave me
> > messages like:
> > access("/var/tmp/sqlite_MyjXYCDJGFYkfnc-journal", F_OK) = -1 ENOENT (No
> such
> > file or directory)
> >
> > However, I have some debug statements in my code which indicate that the
> > failure is during insertion directly in the Main Table (Commit phase of
> that
> > transaction).
> > Any ideas?
> > Thanks!
>
> Do you have write permission on the directory containing the
> database file? That is necessary in order to write to the
> database.
> --
> D. Richard Hipp <[EMAIL PROTECTED]>
>
>
