The sqlite3.exe program is set up as a utility and maintenance tool, not a production environment and is designed to that end. If you want maximum performance it is not the way to go; instead embed the Sqlite calls inside your application and optimize access. If you are performing ad-hoc DB tasks then it or one of the many similar function Ssqlite tools are appropriate. One is the Firefox plug in.
You can imagine that having to compile the SQL for over and over instead of storing and re-using the compiled code adds considerably to overhead on frequently run jobs.. JS Christophe Leske wrote: > John, > > thanks for your suggestions, but i am wondering if any of your > suggestions can be applied to the sqlite3.exe command line application? > > > >>Prepare your statements only once and then use bind. Do not use >> > > How is this done? Can the command line executable be modified in such a way? > > >>sqlite3_exec. Do not open and close the DB for each read, instead open >>once and let the cache work. > > That´s already done. > >>Avoid row scans by defining indices. > > Already there. > >> Use >>the new index selection functionality to force the use of the best >>index. > > Would you care to provide more information about this? A simple link > would be enough. > > >>Place large and less frequently accessed columns at the end of >>the Sqlite rows. >> > > Ok, thanks. Will do. However, i read out the whole row all the time. > BTW, does it help to specify which coloumns i would like instead of all? > I would like to get all minus one coloumn in general. > >>When you obey these rules you will get very good read perfprmance from >>Sqlite. The cacheing is important if you are using a slow disk or flash >>memory. Look at shared cache mode if you have multiple users > > Nope, just one from DVD. > > Thanks again, > _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
