Be aware that the backup process will need to restart from zero after
each write!
Now, if you can setup some kind of IPC between your two processes, then
you could have the update process update the disk base and send
identical data to the reader process, so the latter can update a memory
First of all there's no multi-process equivalent of read_uncommitted.
There's just no way to implement that. If you want such type of
behavior with your current application structure then you should use
some other RDBMS.
For SQLite though I'd suggest you to use some variation of you 3rd
approach:
I have two independent processes that open the same sqlite database.
One inserts data into a table and the other reads from the same table
and creates reports.
The reader process does something like "select * from table" and then
does some heavy processing between each call to sqlite3_step(). It
3 matches
Mail list logo
