Efficient regex scanning


#1

Hello there,

I wan’t to extract all the words from a file and so i wrote the
following code:

file = ARGV[0]
File.open(‘output’,‘w’) {|f|
IO.read(file).scan(/\w+/).each{|w| f.print w}
}

The problem with this code is that it stores all the words in an array
which is not so good in terms of efficiency.
Is there a better way to do it?
Something like IO.read(file).each_scan { foo }

Thanks
Christos


#2

Trochalakis C. wrote:

The problem with this code is that it stores all the words in an array
which is not so good in terms of efficiency.
Is there a better way to do it?
Something like IO.read(file).each_scan { foo }

Thanks
Christos

Scan takes a block form:

ri String.scan

    IO.read(file).scan(/\w+/) {|w| f.print w}

Cheers


Ola B. (http://ola-bini.blogspot.com)
JRuby Core Developer
Developer, ThoughtWorks Studios (http://studios.thoughtworks.com)

“Yields falsehood when quined” yields falsehood when quined.


#3

Trochalakis C. wrote:

Hello there,

The problem with this code is that it stores all the words in an array
which is not so good in terms of efficiency.
Is there a better way to do it?
Something like IO.read(file).each_scan { foo }

Thanks
Christos

Does just using a block with scan do what you need?

IO.read(file).scan(/\w+/) { |word| f.print word }

http://www.ruby-doc.org/core/classes/String.html#M000827

best,
Dan


#4

Hi –

On Wed, 6 Jun 2007, Trochalakis C. wrote:

The problem with this code is that it stores all the words in an array
which is not so good in terms of efficiency.
Is there a better way to do it?
Something like IO.read(file).each_scan { foo }

You could do something like this (untested, and reversing your logic
somewhat):

File.open(file).each {|line| f.print(line.scan(/\w+/)) }

(You might want to join them with a space or something so they don’t
all run together.)

David


#5

On 06.06.2007 13:08, removed_email_address@domain.invalid wrote:

File.open(‘output’,‘w’) {|f|

File.open(file).each {|line| f.print(line.scan(/\w+/)) }

(You might want to join them with a space or something so they don’t
all run together.)

You’re not closing the IO. I know it’s not an issue for a small script
but…

I’d do this:

ARGF.each {|line| puts line.scan /\w+/}

:slight_smile:

Kind regards

robert


#6

Hi –

On Wed, 6 Jun 2007, Robert K. wrote:

file = ARGV[0]
somewhat):

File.open(file).each {|line| f.print(line.scan(/\w+/)) }

(You might want to join them with a space or something so they don’t
all run together.)

You’re not closing the IO. I know it’s not an issue for a small script
but…

It’s not a complete script; I was only showing one line. At the very
least it’s not going to run unless you assign something to f :slight_smile:

David


#7

On Jun 6, 2:00 pm, Ola B. removed_email_address@domain.invalid wrote:

ri String.scan

    IO.read(file).scan(/\w+/) {|w| f.print w}

Cheers

Thanks a lot!
I suppose should have checked first :slight_smile:


#8

Trochalakis C. wrote:

The problem with this code is that it stores all the words in an array
which is not so good in terms of efficiency.
Is there a better way to do it?
Something like IO.read(file).each_scan { foo }

Here’s a thought. Note that it doesn’t handle //m regexen. Like David’s
and Robert’s solutions, it doesn’t read the whole at once. (I guess one
could check for pat.options&Regexp::MULTILINE, and read the whole IO in
that case.)

class IO
def scan pat
if block_given?
each {|line| line.scan(pat) {|s| yield s} }
else
read.scan(pat)
end
end
end

File.open(filename) do |f|
f.scan(/\w+/) {|word| puts word}
end