[lnkForumImage]
TotalShareware - Download Free Software
Confronta i prezzi di migliaia di prodotti.
Asp Forum
 Home | Login | Register | Search 


 

Forums >

comp.lang.ruby

catting files

Mark Probert

2/28/2005 8:17:00 PM


Hi ..

There is approximately an order of magnitude difference in the performance of
these two snippets. Is there a faster way of doing the pure Ruby version?


### use the system call 'cat'
def sys_cat
of = "plato.txt.cat"
clean(of)
l = `cat Plato/*.txt > #{of}`
end


### open each file and copy it
def rby_cat
of = "plato.txt.rby"
clean(of)
off = File.new(of, "w+")
Dir["Plato/*.txt"].each do |f|
text = IO.readlines(f)
off.puts text
end
off.close
end

user system total real
sys_cat 0.000000 0.015625 0.117188 ( 0.167577)
rby_cat 0.937500 0.085938 1.023438 ( 1.064247)


Thanks,

--
-mark. (probertm at acm dot org)
6 Answers

Florian Gross

2/28/2005 8:23:00 PM

0

Mark Probert wrote:

> There is approximately an order of magnitude difference in the performance of
> these two snippets. Is there a faster way of doing the pure Ruby version?
>
> ### open each file and copy it
> def rby_cat
> of = "plato.txt.rby"
> clean(of)
> off = File.new(of, "w+")
> Dir["Plato/*.txt"].each do |f|
> text = IO.readlines(f)
> off.puts text
> end
> off.close
> end

def ruby_cat()
of = "plato.txt.ruby"
clean of
File.open(of, "w") do |off|
Dir.glob("Plato/*.txt") do |f|
off << File.read(f)
end
end
end

You might get better performance by reading the files in 4096 byte
blocks or something similar.

Robert Klemme

2/28/2005 8:29:00 PM

0


"Florian Gross" <flgr@ccan.de> schrieb im Newsbeitrag
news:38hcsvF5n2uqqU1@individual.net...
> Mark Probert wrote:
>
>> There is approximately an order of magnitude difference in the
>> performance of these two snippets. Is there a faster way of doing the
>> pure Ruby version?
>>
>> ### open each file and copy it
>> def rby_cat
>> of = "plato.txt.rby"
>> clean(of)
>> off = File.new(of, "w+")
>> Dir["Plato/*.txt"].each do |f|
>> text = IO.readlines(f)
>> off.puts text
>> end
>> off.close
>> end
>
> def ruby_cat()
> of = "plato.txt.ruby"
> clean of
> File.open(of, "w") do |off|
> Dir.glob("Plato/*.txt") do |f|
> off << File.read(f)
> end
> end
> end
>
> You might get better performance by reading the files in 4096 byte blocks
> or something similar.

.... and binary possibly helps, too.

def stream_copy(in, out)
while ( b = in.read(4096) )
out.write b
end
end

def ruby_file_cat(in, out)
File.open(in, "rb") do |i|
File.open(out, "wb") {|o| stream_copy(i, o)}
end
end

Or similar.

robert

Alexander Kellett

2/28/2005 8:37:00 PM

0

IO.read?
gets(nil)?

On Feb 28, 2005, at 9:17 PM, Mark Probert wrote:
>
> Hi ..
>
> There is approximately an order of magnitude difference in the
> performance of
> these two snippets. Is there a faster way of doing the pure Ruby
> version?
>
>
> ### use the system call 'cat'
> def sys_cat
> of = "plato.txt.cat"
> clean(of)
> l = `cat Plato/*.txt > #{of}`
> end
>
>
> ### open each file and copy it
> def rby_cat
> of = "plato.txt.rby"
> clean(of)
> off = File.new(of, "w+")
> Dir["Plato/*.txt"].each do |f|
> text = IO.readlines(f)
> off.puts text
> end
> off.close
> end
>
> user system total real
> sys_cat 0.000000 0.015625 0.117188 ( 0.167577)
> rby_cat 0.937500 0.085938 1.023438 ( 1.064247)
>
>
> Thanks,
>
> --
> -mark. (probertm at acm dot org)

Javier Valencia

2/28/2005 9:01:00 PM

0

Mark Probert wrote:

>Hi ..
>
>There is approximately an order of magnitude difference in the performance of
>these two snippets. Is there a faster way of doing the pure Ruby version?
>
>
>### use the system call 'cat'
>def sys_cat
> of = "plato.txt.cat"
> clean(of)
> l = `cat Plato/*.txt > #{of}`
>end
>
>
>### open each file and copy it
>def rby_cat
> of = "plato.txt.rby"
> clean(of)
> off = File.new(of, "w+")
> Dir["Plato/*.txt"].each do |f|
> text = IO.readlines(f)
> off.puts text
> end
> off.close
>end
>
> user system total real
>sys_cat 0.000000 0.015625 0.117188 ( 0.167577)
>rby_cat 0.937500 0.085938 1.023438 ( 1.064247)
>
>
>Thanks,
>
>
>

Try mine:

def rby_cat
File.open("file.rby", "w+") do |file|
Dir["*.txt"].each do |f|
file.write(IO.readlines(f))
end
end
end

Austin Ziegler

2/28/2005 9:02:00 PM

0

On Tue, 1 Mar 2005 05:17:03 +0900, Mark Probert <probertm@acm.org>
wrote:
> There is approximately an order of magnitude difference in the
> performance of these two snippets. Is there a faster way of doing
> the pure Ruby version?

> ### use the system call 'cat'
> def sys_cat
> of = "plato.txt.cat"
> clean(of)
> l = `cat Plato/*.txt > #{of}`
> end

def faster_cat
of = "plato.txt.fct"
clean(of)
File.open(of, "wb+") do |outf|
Dir["Plato/*.txt"].each do |inf|
outf.puts IO::read(inf)
end
end
end

As I don't have the "Plato/*.txt" files, I can't test it, but that
you're not splitting the file into multiple lines and iterating over
them will be faster.

-austin
--
Austin Ziegler * halostatue@gmail.com
* Alternate: austin@halostatue.ca

Mark Probert

2/28/2005 9:21:00 PM

0

Hi ..

On Monday 28 February 2005 13:02, Austin Ziegler wrote:
>     of = "plato.txt.fct"
>     clean(of)
>     File.open(of, "wb+") do |outf|
>       Dir["Plato/*.txt"].each do |inf|
>         outf.puts IO::read(inf)
>       end
>     end

Perfect!
user system total real
sys_cat 0.000000 0.007812 0.101562 ( 0.164781)
rb_cat 0.953125 0.070312 1.023438 ( 1.065408)
faster_cat 0.046875 0.078125 0.125000 ( 0.166651)


Thanks, Austin. I knew there was a better way :-)

--
-mark. (probertm at acm dot org)