However even with "vacuum full", the old rows will be removed completely from the storage , but the new rows will always be there with the 'dropped' column still existing under the hood along with the table storage, with just carrying "null" values in it. […] Is this understanding correct?
No. The table rewrite process involves creating new tuples that exactly conform to the current row specification. The potentially non-null data present in live tuples for columns that have been dropped are not copied into the newly constructed tuples.
My bad, stopped at the code comment. Apparently the data is just nulled, not removed, the current row descriptor contains those columns with “is dropped” and since this behavior doesn’t change the catalogs in this way the new ones must as well. We just get the space back.
Thank you for the confirmation.
And if someone wants to fully remove that column from the table , then the only option is to create a new table with an exact set of active columns and insert the data into that from the existing/old table and then rename it back to old. Is this correct understanding?