On Wed, Sep 26, 2012 at 10:59 AM, Christopher Browne <cbbrowne@gmail.com> wrote:
> A different place where I wound up having to jump through considerable
> hoops when doing schema analytics was vis-a-vis identifying functions.
> I need to be able to compare schemas across databases, so oid-based
> identification of functions is a total non-starter. It appears that
> the best identification of a function would be based on the
> combination of schema name, function name, and the concatenation of
> argument data types. It wasn't terribly difficult to construct that
> third bit, but it surely would be nice if there was a view capturing
> it, and possibly even serializing it into a table to enable indexing
> on it. Performance-wise, function comparisons turned out to be one of
> the most expensive things I did, specifically because of that mapping
> surrounding arguments.
pg_proc.oid::regprocedure::text has been pretty good to me for this
sort of thing.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
