Thanks guys. You really helped me. I think I know what I am going to do. You see, as I am using a single connection object there is at least one selection statement executed at the beginning of each process. Then, during the program lifecycle there could be a few updates executed, that use the same connection object. So, the transaction that was opened with a SHARED lock would be converted to a RESERVED lock until committed.
Therefore, I am thinking of using separate (local) connections for the routines that do the updates, and leaving the global connection object for the selects only. I have estimated that there are not a lot of places in the program what do SELECT and then UPDATE, so, for these routines I will be doing the following: 1. open new connection 2. execute BEGIN IMMEDIATE 3. do SELECT/UPDATE/DELETE (no change) 4. do the COMMIT (or ROLLBACK in case of an exception) 5. close connection. For the rest of the program, I will leave the global connection object. Andrew Cherednik _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
