MaximG for only the top 200 companies if they had 1 billion...

MaximG for only the top 200 companies if they had 1 billion shares each you would not only need 200,000,000,000 unique records to identify each share. That's one table. You would then need the history for each shares trades back to when the share was issued. New transactions accruing at the rate of the volume of shares traded each day on the ASX. The ASX obviously just counts them. We are talking a new row each time and individual share is traded. Then we could work out which shares were held for 1 minute and which shares were held for 40 years.

How many Petabytes are we talking. Does even TeraData deal in several petabytes yet?