-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 05/02/12 10:13, Hamish Allan wrote: > 1) Is concatenation on UPDATE performed in-place?
The values for a row are stored sequentially. Changing the size of a value will at least require rewriting the row. I suggest reading the documentation on the file format which will make it clear how everything is stored: http://www.sqlite.org/fileformat2.html I also suggest you use a profiler to work out where your bottlenecks actually are rather than worrying about theoretical "efficiency" like this. (Put another way, if your code spends 3% of its time in the critical path doing these updates then look in the other 97% for improvements.) If you are trying to reduce the amount of I/O then you'll need something repeatable while you tweak and tune. blktrace on Linux can help see what I/O happens. If you have more deep instrumenting in your program then you can make a new VFS that calls the existing ones but in the xRead and xWrite routines also grab the current query being executed so you can assign blame. TLDR: Only worry about "efficiency" when you have accurately measured issues, and then make sure you have instrumentation as you fine tune to improve it. This is rarely a good use of your time. Roger -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) iEYEARECAAYFAk8u1FwACgkQmOOfHg372QTHbgCfRl1sY63frasmccFBckdwRWjg RPAAoJ9xECJNVMu+jvGnbOpOuLZQ0VfQ =Zuhs -----END PGP SIGNATURE----- _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
