Go to file
Worros 51ce6fc191 Index mods '+0' from sqlcoder on mailing list
Have been tested and provide a significant improvement on import.

"""
Along with new indexes, I think some of the sql in fpdb_simple can be
altered to make it use existing indexes more effectively. Generally
these mods affect statements where the db has a choice of 2 indexes to
use and I think it is choosing badly, e.g. it is using tourneyTypeId
which could match the whole table when it could use the new
siteTourneyNo index which will return very few rows.

To stop the db from using an index I'm just adding zero to the field, so
something like
   playerId = %s
changes to
   playerId+0 = %s
We humans can tell this is no real change at all, but the db sees this
as a change from <field> = <value> to <expression> = <value>, and it can
only use the index if the <field> the index is on is being tested.
(Nowadays you can index expressions in some databases but we'll ignore
that!)
"""
2008-12-05 17:18:13 +09:00
docs Merge branch 'master' of git://git.assembla.com/fpdb.git 2008-11-10 03:24:05 -05:00
ignore-me_perl6 windows git is buggy 2008-11-10 03:22:21 -05:00
packaging p142 - commented tourney summary parsing as it fails on me. this is alpha9 2008-11-10 02:02:12 +00:00
pyfpdb Index mods '+0' from sqlcoder on mailing list 2008-12-05 17:18:13 +09:00
regression-test windows git is buggy 2008-11-10 03:22:21 -05:00
utils p92 - patch from carl to fix utils/get_db_stats 2008-09-22 03:34:28 +01:00
website Updated Grapher screenshot 2008-11-07 18:35:30 +13:00
create-release.sh p90 - release script renames HUD_config.xml.example so user doesnt have to 2008-09-21 23:38:22 +01:00
prepare-git.sh git19 (REIMPORT needed) - updated everything to use new action counting method -> it half works, but fails to store (or print) a substantial proportion of the action_nos 2008-08-08 22:03:43 +01:00