OpenEuphoria

Re: Fast appending and sorting of alot of short strings

• Posted by Pete Lomax <petelomax at blueyonder.co.uk> Oct 09, 2004
• 529 views

On Sat, 09 Oct 2004 05:47:34 -0700, Tommy Carlier
<guest at RapidEuphoria.com> wrote:

>Here's a piece of code from Win4Eu 

I applied my theory to it and compared your:
>}}}
<eucode>
>global function addToCollection(sequence collection, object element)
>	-- Add an element to a collection
>	integer count, growSize
>	count = collection[length(collection)] + 1
>	growSize = count - length(collection)
>	if growSize > 0 then
>		if growSize > GROWSIZE then
>			collection &= repeat(0, growSize + GROWSIZE)
>		else collection &= BUFFER
>		end if
>	end if
>	collection[count] = element
>	collection[length(collection)] = count
>	return collection
>end function
></eucode>
{{{

with:

sequence collection

procedure addTocollection(object element)
	-- Add an element to a collection
	integer count, growSize
	count = collection[length(collection)] + 1
	growSize = count - length(collection)
	if growSize > 0 then
		if growSize > GROWSIZE then
			collection &= repeat(0, growSize + GROWSIZE)
		else collection &= BUFFER
		end if
	end if
	collection[count] = element
	collection[length(collection)] = count
end procedure

They are identical apart from one is a function acting on and
returning a parameter, whereas the other is a procedure acting
directly on a file-level variable.

It may surprise you to find that it is 66 times faster.

The testing code I used to arrive at that figure was:

sequence MyTable
	collection=newCollection()
	MyTable=newCollection()
integer c1,c2
atom t
	t=time()+1
	c1=0
	while t>time() do
		c1+=1
		addTocollection(c1)
	end while
	t=time()+1
	c2=0
	while t>time() do
		c2+=1
		MyTable=addToCollection(MyTable,c2)
	end while
?c1
?c2
?c1/c2
abort(0)

Of course it is a pain to have to duplicate code for each table, but
for that kind of gain it is well worth it.

Regards,
Pete