On 3 Sep 2011, at 4:27pm, Mohit Sindhwani wrote: > for our data, we can get savings in the region of 25% - 33% in the case of > strings being stored in a language that does require 3bytes/ character. So, > given that, we should explore UTF-16 in more detail. However, we also have a > lot of text that is only in English - so, it seems that we should go down the > path of separating the data in the two languages and use an ATTACH to bring > in the other language. That may be best for our needs.
Have you tried speed tests on your platform ? It's hard to tell which will be faster because it depends on what language and OS you're using that interacts with SQLite. So if you have your schema designed and any part of your application written you could try doing lots of information shuffling and see if UTF-8 or UTF-16 is usefully faster than the other. Saving space for your database files may be interesting, but it's useful only if you're short of space, or need faster backups, or something else related. These days most work computers have tons of free space on their hard disks. Simon. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
