Josh Berkus <josh@agliodbs.com> writes:
> Tom,
>> To put it more bluntly: exactly what are you accomplishing here that
>> isn't already accomplished, in a *truly* standard fashion, by the
>> INFORMATION_SCHEMA? Why do we need yet another nonstandard view on
>> the underlying reality?
> To quote myself:
> Q: Why not just use information_schema?
> A: Because the columns and layout of information_schema is strictly defined by
> the SQL standard. This prevents it from covering all PostgreSQL objects, or
> from covering the existing objects adequately to replicate a CREATE
> statement. As examples, there is no "types" table in information_schema, and
> the "constraints" table assumes that constraint names are universally unique
> instead of table-unique as they are in PG.
So? If you want reality, look at the catalogs.
I think that in a release or three, these views will be just as
distorted a representation of the underlying reality as the
information_schema is now. Either that or you'll be changing them
incompatibly. You can't have both truth and a greater degree of
stability than the underlying catalogs.
So my opinion remains "what's the point?". All you have really
accomplished is some editorialization on table/column names.
regards, tom lane