OpenEuphoria

Re: Looking for the best "Parse"...

• Posted by Igor Kachan <kinz at peterlink.ru> Oct 27, 2001
• 385 views

Ýòî ñîîáùåíèå â ôîðìàòå MIME ñîñòîèò èç íåñêîëüêèõ ÷àñòåé.

------=_NextPart_000_01C15F3C.6EFF8300

Hi Derek, Hi Jiri, 

I have prepared new variant of this parse play.
I apply the "delete double" method to jparse2()
function.

This new iteration is in the attachment 
with my own bench.

You can see my results below.
These are times of 10000 iterations  on 
my a 386 in plain Dos-32 mode

jparse2()            jparse3() -- new

21.59                21.48
21.59                21.48
21.64                21.47
21.59                21.47
21.59                21.47
21.58                21.48
21.58                21.53
21.59                21.53
21.59                21.47
----------------------------------average
21.59                21.48

Note, I don't know algorithms well enough,
both Cassidy's and any new, I just
read about that fox, which jumps over lazy dog,
to see if there is misprint in my new
iterations to the best function or no there is not.
I am not real author of any task and solution
in this play, I just search for doubled
things and try delete them with the help of 
Euphoria's inner power, so my name and these tasks
and functions are not well compatible things,
I think. And "delete double" method is an old one,
same as ... I don't know as. 

If you'll see Colin's solution near the modifired
Cassidy's function, you can see Colin just
apply "double delete" method to that recursive
solution.

You can also to apply this "delete double" method 
to Cassidy's *test frase* and delete mistakes from
that classic "delete double" test   

OK to finish this subject for the robast 
Eu Dos-32 programming ?

I say, there is OK with humor here, no ?

Regards,
Igor Kachan
kinz at peterlink.ru








------=_NextPart_000_01C15F3C.6EFF8300
Content-Type: application/octet-stream; name="Parse1.e"
Content-Transfer-Encoding: 7bit
Content-Description: Parse1.e (E )
Content-Disposition: attachment; filename="Parse1.e"


include wildcard.e

 function jparse2(sequence s, sequence s1, sequence s2)
     -- replace every occurrence of s1 in s by s2
     -- case insensitive search
     sequence os,u,v
     integer i,j,m

     os = {}
     u = lower(s)
     v = lower(s1)
     m = length(s1)
     i = match(v, u)
     j = 1
     while i do
	 os &= s[j..j+i-2] & s2
	 j += i+m-1
	 u = u[i+m..length(u)]
	 i = match(v, u)
     end while
     return os & s[j..length(s)]
 end function

integer k,i,j,m
sequence os,u,v

 function jparse3(sequence s, sequence s1, sequence s2)
     -- replace every occurrence of s1 in s by s2
     -- case insensitive search
--     sequence os,u,v
--     integer k,i,j,m

     os = {}
     u = lower(s)
     v = lower(s1)
     m = length(s1)
     j = 1
     for t=1 to 111111111 do
       i = match(v, u) if not i then exit end if   
	   k = i+m
	 os &= s[j..j+i-2] & s2
	  j += k-1
	   u = u[k..length(u)]
     end for
     return os & s[j..length(s)]
 end function

sequence text

text = jparse3 ("The $A jumped over the lazy dog.","$A","quick brown fox")

puts (1, text & "\n")


atom T

T=time()
for p=0 to 10000 do
text = jparse2 ("The $A jumped over the lazy dog.","$A","quick brown fox")
end for
T=time()-T
? T

T=time()
for p=0 to 10000 do
text = jparse3 ("The $A jumped over the lazy dog.","$A","quick brown fox")
end for
T=time()-T
?T


------=_NextPart_000_01C15F3C.6EFF8300--