Reece Hart wrote:
On Tue, Mar 8, 2011 at 5:52 PM, Darren Duncan <darren@xxxxxxxxxxxxxxxx>wrote:
Try starting with an enum type to define the possible values:
CREATE TYPE Consequence_Type
AS ENUM ('ESSENTIAL_SPLICE_SITE','STOP_GAINED',
'STOP_LOST','COMPLEX_INDEL','SPLICE_SITE');
... and then you could try using "ARRAY OF Consequence_Type" or some such.
Arrays occurred to me, but they don't give a set (i.e., a consequence type
can appear more than once) unless I write the code to dedupe the array.
However, if I were going to put that much effort into it, I might as well
represent the set directly and obviate the dedupe. Or, am I missing
something from your suggestion?
You also have another option, which is to split consequence_type into
a second table, and have a record in there for each value in the set. You would
maintain a lack of duplicates by having an ordinary unique/primary key
constraint on the second table. This second-table approach is logically
equivalent to your use of a set-valued field. -- Darren Duncan
--
Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
