Sort pages of an OCR scan pdf, by 2 fields that are the bottom of each page

As long as there is a clean OCR, a script can be written to aquire the page text using the "doc.getPageNthWord()" function. The page text should be concatonated into a large string, which can  be searched using  regular expressions for the key elements needed for sorting. I would suggest creating an array of the search elements, so there is an entry for each page. The entries could be an object containing the original page number and the search elements.  The array can then be sorted, giving the new locations where the pages are to be moved. 

I would assume the "# of #" text is used to group pages, where as the date is the real sorting element. Is this correct?

This is not a trivial operation. I've done this sort of thing many times.