How can I replace just the corrupt page(s) in the domain file of iWeb using Time Machine?

Question

I back up with Time Machine and have an extensive elaborate website I created in iWeb '09 over a couple months and publish to a local folder and then upload to my server, but in the last few days I notice certain pages (that I haven't even worked on or touched) somehow become corrupted or "cross-contaminated" with elements and images from other pages.  Once I see they are corrupt I make sure not to publish them (if the current published versions are the correct, non-corrupt versions) or if they did get published I can use Time Machine to retrieve the .html file and page files folder for that specific page and replace it in my published folder/server so it shows correctly on the Web.  However, that does not replace the corrupted page(s) you see and work with when you launch iWeb and try to edit or continue working on that page.
I am confused as to how I go to the package contents of the domain file and replace just that page with a previous version from time machine.  I don't want to replace the entire domain file because I have new changes I made to other pages even in the past hours.  How can I keep the good pages and just get earlier, non-corrupted versions of the corrupt page(s)? I know it's not as easy as with the published site folder where you can just replace the page's .html file and folder, but I don't want to have to re-create the corrupted page(s) from scratch or have to replace it with the last non-corrupted domain file and then have to redo all my recent changes to new pages I made before discovering the corrupt page(s).  Thanks for your help as now I can't make changes to the page(s) within iWeb itself.

Answer 1

Thanks for the response, Wyodor.  I don't know what that is, but I'll have a look.  Is it an alternative to iWeb or a way to transfer pre-existing iWeb sites?  When you say merge domain files, is that like so multiple copies of the same site show up then you can pick and choose the non-corrupted pages and group them then delete the corrupted ones?  I am on Snow Leopard with no plans or need to upgrade anytime soon.
And yes, I will read your links but just wanted to ask those questions.  Maybe they'll answer my questions, maybe not.
I was able to discern that within the domain file is a domain folder with all the site folders, each with their own page ".gz" files which expand into ".xml" files.  I was trying to figure out if you could simply drag the corrupt pages out that way and replace them with backup copies that are still good.  I am having trouble discerning which pages are which as they all have random names like site-page-30F175E3-AE33-4F10-A490-1A096D9B185B.xml and although I expanded and opened each in Text Wrangler, I still couldn't discern which were which for sure, and trial and error proved cumbersome.  Also, I did notice some of the later corrupted domain files had one or two more pages than the site itself has, so not sure how they got added or duplicated or what.
Again, I'll look at your links, but do you know about swapping out individual page .xml files this way within the domain file?