more compact storage?

for c.tx.storage:

• 40M rows
• 4.7M distinct values
• 20M null values

-> 4.7M distinct values for 20M rows. It might be possible to gain a great amount of space by storing those values in a separate table.

v970=# select count(distinct(storage)) from c.tx ;
  count  
---------
 4728858
(1 row)

v970=# select count(*) from c.tx  where storage is null ;
  count   
----------
 20597271
(1 row)

v970=# select count(*) from c.tx ;
  count   
----------
 40018845
(1 row)