Forum Moderators: phranque
I've read the thread Changing dynamic to static URLs [webmasterworld.com], and some of the articles on rewrites here. That is essentially what I need to do. But there is one twist that I can't quite figure out.
In the thread, the example seems to assume your parameters will always be in the same order:
http://example.com/mypage.html?p1=v1&p2=v2&p3=v3
However, in my case, there is no such guarantee. So what I need to do is that for all of the following (for example):
http://example.com/mypage.html?p1=v1&p2=v2&p3=v3
http://example.com/mypage.html?p2=v2&p3=v3&p1=v1
http://example.com/mypage.html?p1=v1&p3=v3&p2=v2&
http://example.com/v1/v2/v3/mypage.html
Does this mean I need some kind of chain of RewriteRules? I can do this if I am only picking out one parameter, but the added parameters seem to put in a bit of twist (which may not be, but I'm new at this).
To add to this, the result may be a mixed URL. That is, I can have:
http://example.com/mypage.html?p2=v2&p3=v3&p1=v1&p4=v4
http://example.com/mypage.html?p4=v4&p2=v2&p3=v3&p1=v1
need to end up as:
http://example.com/v1/v2/v3/mypage.html?p4=v4
That is, for those remaining parameters I'm not picking out, they must remain as part of the query (and those that I've picked out need to be removed from the query).
Is this possible with the mod_rewrite? Or really should be looking at a completely different solution?
Thanks!
Vince
[edited by: jdMorgan at 4:19 am (utc) on July 21, 2007]
[edit reason] exampe.com [/edit]
i would internally rewrite the url to a "canonical" form:
http://www.example.com/mypage.html?p1=v1&p2=v2&p3=v3&otherparams=othervalues
http://www.example.com/v1/v2/v3/mypage.html?otherparams=othervalues
Here's what I'd try:
RewriteCond %{QUERY_STRING} ^(([^&]+&)*)p1=([^&]+)((&[^&]+)*)&?$
RewriteRule ^mypage\.html$ %3/mypage.html?%1%4 [C]
RewriteCond %{QUERY_STRING} ^(([^&]+&)*)p2=([^&]+)((&[^&]+)*)$
RewriteRule ^([^/]+)/mypage\.html$ $1/%3/mypage.html?%1%4 [C]
RewriteCond %{QUERY_STRING} ^(([^&]+&)*)p3=([^&]+)((&[^&]+)*)$
RewriteRule ^([^/]+)/([^/]+)/mypage\.html$ http://www.example.com/$1/$2/%3/mypage.html?%1%4 [R=301,L]
Jim
i'd like to change my answer to what he said...
May be just my programmer background but I tried the implementation like this, using env var:
(recap, more precisely, mapping
http://example.com/search/?p2=v2&p1=v1&p4=v4&p3=p3&
http://example.com/search/?p1=v1&p4=v4&p3=p3&p2=v2
http://example.com/search/?p1=v1&p2=v2&p3=p3&p4=v4
all to
http://example.com/search/p1/p2/p3?p4=v4
#-- process only the ones start with "/search", skip for
#-- other types
RewriteCond %{REQUEST_URI}!^/search[/]*$ [NC]
RewriteRule .? - [S=4]#-- extract the parameters
RewriteCond %{QUERY_STRING} ^(([^&]+&)*)p1=([^&]+)&?(([^&]+&?)*)&?$ [NC]
RewriteRule .? - [E=pV1:%3,E=myQS:%1%4]RewriteCond %{ENV:myQS} ^(([^&]+&)*)p2=([^&]+)&?(([^&]+&?)*)$ [NC]
RewriteRule .? - [E=pV2:%3,E=myQS:%1%4]RewriteCond %{ENV:myQS} ^(([^&]+&)*)p3=([^&]+)&?(([^&]+&?)*)$ [NC]
RewriteRule .? - [E=pV3:%3,E=myQS:%1%4]#-- reassemble in single redirect
RewriteRule . [%{HTTP_HOST}...] [R=301,L]
This seem to work and seems cleaner and clearer. Basically I'm extracting the parameters I need, then reassemble for redirect. If I'm writing a program, this might be what I'd do. Also may have some "default" variable set, like:
#-- default value
RewriteRule .? - [E=pV3:defaultV3]#-- extract the parameter
RewriteCond %{ENV:myQS} ^(([^&]+&)*)p3=([^&]+)&?(([^&]+&?)*)$ [NC]
RewriteRule .? - [E=pV3:%3,E=myQS:%1%4]
Now, a couple more questions:
1) Is there any reason for choosing the first way (with chaining) versus the env variable? Like performance concerns (as I'm dealing with a potentially busy site)?
2) I notice in the "catch all" RewriteRule, sometimes "." is used and sometimes ".?" is used. I understand the regex as one being any one char (.) and the other could be nothing at all? (.?) Are there reasons for using one over the other (where in our case here we really isn't matching rather than saying (apply this rule)?
Thanks!
Vince
[edited by: jdMorgan at 6:41 pm (utc) on July 26, 2007]
[edit reason] Disable BBcode smilies [/edit]
2)
The "." pattern means at least one character, but since it's un-anchored it will match one or more characters.
The ".*" pattern means match zero or more characters.
Examining the context of the rules you're referring to --some with "." and some with ".*-- you will find that some have RewriteConds that will be false if there isn't at least one character in the URL-path, while others will match either way. Since RewriteConds are not processed unless the RewriteRule pattern matches (See "Rule Processing" in the Apache docs), we might as well use the RewriteRule pattern to avoid parsing the RewriteCond(s) if there is no chance that the RewriteCond(s) will match.
Regarding your exact implementation, Id say that
RewriteCond %{QUERY_STRING} ^(([^&]+&)*)p1=([^&]+)((&[^&]+)*)&?$ [NC] RewriteCond %{QUERY_STRING} ^(([^&]+&)*)p1=([^&]+)&?(([^&]+&?)*)&?$ [NC] Jim
[edited by: jdMorgan at 6:40 pm (utc) on July 26, 2007]
