Dear SQLite Community, Please help me with your expertise and advice.
Environment: - Windows service which reads/updates any one of 60 different settings (integer values and strings) stored in SQLite database. This is .db file on disk. Another windows service (over which I have not control) also updates this .db files once in while.So SQLite DB is primary settings store. Some Analysis: ProcMon displays that there are 80-90 'ReadFile' requests per transaction > on my application. The transaction finishes in 1-3 seconds and user can at start at most 15 > transactions at any given instant. Again the frequency of transactions is > not rigorous, it is at best intermittent. Say at max, 10 transactions in 2 > mins. Now I think each ReadFile shown in ProcFile will not hit the disk. How can improve the service performance here? Keeping in mind, - There will be no new inserts (ignoring addition of few hundred URLs once a day while overwriting the older ones). Size of the DB will be more or less constant in range 1-2MB or at most 5 MB. - Most of the times service is reading values. - Another service update some of these settings. (say once an hour) Intention: Super fast response to the user. Is this case good candidate for using SQLite In-Memory database? If not, how can I reliably know the number of hits to the SQLite .db file on the disk? Thank you.
