In a current project, I would like to store small binary data in some fields of a database. PostgreSQL does not support the BINARY or VARBINARY types, and it appears that VARCHAR is not safe for \000s (and lexically sorts anyhow), so I have been using BYTEA. Unfortunately, PostgreSQL 7.0.3 lacks the operations necessary to build indices on BYTEA types. The included patch implements the missing operations. I wasn't able to find any guidelines on choosing OIDs, other than "don't choose ones already in use or obviously reserved", so I have made use of the range 1900-1916. If there isn't a better way to do what I'm trying to do, I'd appreciate seeing this in a future PostgreSQL release. --Jered Floyd jered(at)permabit(dot)com
Attachment:
pgsql-7.0.3-byteaops
Description: patch against PostgreSQL 7.0.3 to implement bytea_ops