-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 05/26/2011 10:41 PM, Frank Chang wrote: > > Roger Binns, Thank you for your reply.
You never answered the important bit - is your concern about initial population of the database or about runtime later on. > Would you expect us to get faster sqlite WAL reads without the wal > checkpoint after commit? Is it possible there is another method for obtaining > get faster sqlite WAL reads ? WAL is mainly a way of helping with writes not reads. In any event doing random things and hoping they are the most performant is not a good way of going about this. Make a benchmark that most closely matches your data and access patterns and that can be run repeatedly, Once you have that you can try with and without WAL, different page sizes, different transactions strategies etc. Roger -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) iEYEARECAAYFAk3fQ5kACgkQmOOfHg372QR7+gCfZZ5Zaw9l1TMSH66Zl3NJCJBn IswAoNd/Ci6IGvb8Yruino61mVEmIIo5 =JQAT -----END PGP SIGNATURE----- _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
