I think we don't want to merge this branch into lp:zeitgeist until we
understand the whole consequences of this change.
--
https://code.launchpad.net/~thekorn/zeitgeist/fix-672965-opt_timerange_queries/+merge/40412
Your team Zeitgeist Framework Team is requested to review the proposed merge of
l
Running ANALYZE after inserting the sample data and before running the
find_event queries does not change anything, for both, this branch and
lp:zeitgeist.
I must admit I don't understand why not using the timestamp index is
*significantly* faster, but it is (0.26s -> 0.07s, 0.20s -> 0.06s etc.)
A few comments:
* I am not convinced that the perceived improvement is statistically
significant. I haven't done the number crunching, but it's a bit fishy...
* I'd really like to understand *why* it's bad to use the time index here and
why sqlite goes on to use it? Does running ANALYZE (rebu
The question which came in my mind after changing the code was: does the
performance of other queries suffer from this change, so I changed the script
attached to the bug to test all ResultType
Before the change:
---> LeastPopularActor: get len(ids)=6 using .find_eventids() in 0.078450s
---> Le
Markus Korn has proposed merging
lp:~thekorn/zeitgeist/fix-672965-opt_timerange_queries into lp:zeitgeist.
Requested reviews:
Zeitgeist Framework Team (zeitgeist)
Related bugs:
#672965 find_event() queries with timerange other than TimeRange.always() are
slow
https://bugs.launchpad.net/bug
5 matches
Mail list logo