On Apr 9, 11:53 am, Jeremy Evans <[email protected]> wrote:
> Well, it can't return self, but it can return a direct clone. Try the
> patch athttp://pastie.org/911966.txtand let me know if it has the
> functionality you had in mind.
Unfortunately, there's a 15% performance decrease for Dataset#filter
using that patch, so I can't accept it:
With:
irb(main):019:0> puts
Benchmark.measure{100000.times{ds.filter(:a=>1)}}
3.600000 0.450000 4.050000 ( 4.058307)
=> nil
irb(main):020:0> puts
Benchmark.measure{100000.times{ds.filter(:a=>1)}}
3.690000 0.340000 4.030000 ( 4.028950)
=> nil
irb(main):021:0> puts
Benchmark.measure{100000.times{ds.filter(:a=>1)}}
3.620000 0.440000 4.060000 ( 4.054360)
Without:
irb(main):003:0> puts
Benchmark.measure{100000.times{ds.filter(:a=>1)}}
3.210000 0.270000 3.480000 ( 3.481032)
=> nil
irb(main):004:0> puts
Benchmark.measure{100000.times{ds.filter(:a=>1)}}
3.190000 0.260000 3.450000 ( 3.454075)
=> nil
irb(main):005:0> puts
Benchmark.measure{100000.times{ds.filter(:a=>1)}}
3.070000 0.400000 3.470000 ( 3.477142)
Thankfully, most of that 15% seems to be overhead of creating objects
unnecessary in every call, which is easy to eliminate (see
http://pastie.org/912025.txt):
irb(main):003:0> puts
Benchmark.measure{100000.times{ds.filter(:a=>1)}}
3.370000 0.200000 3.570000 ( 3.567935)
=> nil
irb(main):004:0> puts
Benchmark.measure{100000.times{ds.filter(:a=>1)}}
3.280000 0.260000 3.540000 ( 3.544439)
=> nil
irb(main):005:0> puts
Benchmark.measure{100000.times{ds.filter(:a=>1)}}
3.220000 0.350000 3.570000 ( 3.573647)
This brings the performance decrease to about 2.5%, which is still
more than I'd like, but I'll at least consider it.
Jeremy
--
You received this message because you are subscribed to the Google Groups
"sequel-talk" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/sequel-talk?hl=en.
