how to make a sparse database

Hi, I am trying to represent a table (like an excel page for example). The project is to crawl the web and pull in the text of webpages and insert all the words into a table. The crawler reads a page and then breaks it up into the list of all the words in the page. I am using beautiful soup and urllib2.
The top row of the table is a non repetitive (no duplicates) list of all the words in all the pages. Each row represents each url and the entry in each column is the number of  times that word has appeared in that particular page (for the particular url that is being   represented by this row). So I will crawl many (maybe 5000) pages. That means that the top row of this table is the list of all the words that are in these 5000 pages. There may be something like 4 or 5 thousand different words. But any particular page and therefore any particular row of this table will probably only contain a small subset (say 400) of these words. Thus this table will be a sparse table or matrix.
I don't know if I should use a dictionary, a shelf, or some other structure to do this. If it is a dictionary would it be a dictionary with the keys being the urls? and the values being a list? Or should the keys be the 4 or 5 thousand different words and values be another dictionary with the keys being the urls and the values being a list of the ???
I am confused. Should I just have a list with the keys being the urls and the values being a list of the frequencies? That sounds like what I want but it does not include the list of all the possible words (the 4 or 5 thousand words)
Any help would be appreciated.

Here is some of the code (it is not mine)

1:
2:
3:
4:
5:
6:
7:
8:
9:
10:
11:
12:
13:
14:
15:

def addtoindex(self,url,soup):
  if self.isindexed(url): return
  print 'Indexing '+url
  # Get the individual words
  text=self.gettextonly(soup)
  words=self.separatewords(text)
  # Get the URL id
  urlid=self.getentryid('urllist','url',url)
  # Link each word to this url
  for i in range(len(words)):
      word=words[i]
      if word in ignorewords: continue
      wordid=self.getentryid('wordlist','word',word)
      self.con.execute("insert into wordlocation(urlid,wordid,location) \
      values (%d,%d,%d)" % (urlid,wordid,i))