Reg Ex help 2

[guitardave78]

Hi all.
I have the following 2 expressions
/.*?-reviews~([^~]*).*\.htm
/.*?-reviews~([^~]*)~([^~]*).*\.htm

These match up to the following strings

/sam-and-max-reviews~10.htm
and
/sam-and-max-reviews~10~PC.htm

These are then used in replaces using $1 $2 etc

Is there a way to combine the expressions

I thought that
/.*?-reviews(?:~([^~]*)){0,3}.*\.htm
May do it, but I only have access to the last match using the $1 group

Any ideas?


}...the bane of my life!

http://www.yetanotherreviewsite.co.uk

http://www.fuzzyd.co.uk

 

[monksnake]

I'm not good at these things, so the best I'll do is try, it'll help me learn. I'm trying to show that you would want the empty string OR the ~([^~]*) part.

/.*?-reviews~([^~]*)[!](()|~([^~]*))[/!].*\.htm

<.

 

[guitardave78]

Mm not quite, has a few to many groups returned!
Like the idea though, oh get hold of expresso regular expression editor, will really help out!!

}...the bane of my life!

http://www.yetanotherreviewsite.co.uk

http://www.fuzzyd.co.uk

 

[monksnake]

Can you show a few examples of what you got returned that you DIDN'T want returned from my example??

<.

 

[monksnake]

I got a new one, not as cryptic since you only need those 2 to match:

/.*?-reviews~10(()|~PC).htm

<.

 

[tsuji]

Like so?
[tt]
dim rx,s,a

set rx=new regexp
with rx
.pattern="^/.*?-reviews((~[^~]*?)+)\.htm$"
.ignorecase=true
'etc etc if needed
end with

s="/sam-and-max-reviews~10~PC~abc~xyz.htm"

if rx.test(s) then
a=split(rx.execute(s)(0).submatches(0),"~")
'a(0) blank, a(1)=10 etc
'monitor the ubound(a) if need restriction
'or replace the "+" by {1,n}
'other manipulations according to need
end if
[/tt]
Slight modifications if allowing no "~..~...." suffix.

 

[guitardave78]

Both interesting, second one should work...but the reason i need to do this with matched groups is that I am not able to use a nice loop, I am using isapi_rewrite, i posted here as I know you guys are good at the reg exp!!

I tried this
/.*?-reviews~([^~]*)(?(~)~([^~]*)).*\.htm$
Which is nice but i need a similar pattern for the match.
I think you both deserve stars for the help!!

}...the bane of my life!

http://www.yetanotherreviewsite.co.uk

http://www.fuzzyd.co.uk

 

[tsuji]

I don't understand exactly the problem. But if you think you have to make good use of $1,...$n, the following shows you how to determine the matching n and each value of them ($1,...$n). The speed of it may not be the best of the world, but it shows how to proceed.
[tt]
set rx=new regexp
with rx
.ignorecase=true
'etc etc if needed
end with

s="/sam-and-max-reviews~10~PC~abc~xyz.htm"

n=1
nmax=100 'n limit to reasonably big according to need
bfound=false

do while (not bfound) and (n<nmax)
spattern="^/.*?-reviews"
for i=0 to n-1
spattern=spattern & "(~[^~]*?)"
next
spattern=spattern & "\.htm$"

rx.pattern=spattern

if rx.test(s) then
bfound=true
response.write "last-n=" & n & "<br />" & "$last-n=" & rx.replace(s,"$" & n) & "<br />"
if n>1 then
response.write "$last-but-one=" & rx.replace(s,"$" & (n-1)) & "<br />"
end if
if n>2 then
response.write "$1$2=" & rx.replace(s,"$1$2") & "<br />"
end if
else
n=n+1
end if
loop
if not bfound then 'or n=nmax at this level
response.write "no match found" & "<br />"
end if
[/tt]