In Redshift, you appear to not be able to drop a column if there is a view dependent on your table, regardless of whether or not your column is referenced.
step 1 - Create table
create table codenames
(
id int identity(0, 1),
name text,
code text
)
step 2 - Create view
create view codenames_names_only
as
select name
from codenames;
step 3 - drop unused column
alter table codenames
drop column code;
At this time, you get the following error:
Error: ERROR: cannot drop table x column code because other objects depend on it
Hint: Use DROP ... CASCADE to drop the dependent objects too.
The above works fine in barebones PostgreSQL (SQLFiddle link)
Thoughts? Ways to get around this without dropping/modifying the views?
