Обсуждение: Drop column and Access

Programs used -Postgres 7.3,SuSE 7.2,psqlODBC 7.02.00.04,Access95,97,Win95,98.
I have a Postgres database running on a Linux server. I dropped a column on
one of the tables. When I tried to relink the table in Access it complained
it couldn't find the table because of a naming violation. The name had stayed
the same so I was somewhat confused. When I looked at the table via
phpPgAdmin I saw the dropped column with a column name of
...pg.dropped.table7...(as well as I can remember). This explained the naming
violation. The only way I could get Access to see the table was to recreate
by storing the nondropped columns in a temp table dropping the original table
and then restoring the table from the temp table. Is there a way of removing
the reference to the dropped column so less enlightened applications don't
have problems?

    Thank you
    Adrian Klaver
    aklaver@attbi.com

Adrian Klaver <aklaver@attbi.com> writes:
> Programs used -Postgres 7.3,SuSE 7.2,psqlODBC 7.02.00.04,Access95,97,Win95,98.
> I have a Postgres database running on a Linux server. I dropped a column on
> one of the tables. When I tried to relink the table in Access it complained
> it couldn't find the table because of a naming violation.

Sounds like psqlODBC may not yet know it's supposed to ignore pg_attribute
entries with attisdropped true.

            regards, tom lane

Tom Lane wrote:
>
> Adrian Klaver <aklaver@attbi.com> writes:
> > Programs used -Postgres 7.3,SuSE 7.2,psqlODBC
> > 7.02.00.04,Access95,97,Win95,98.
> > I have a Postgres database running on a Linux server.
> > I dropped a column on one of the tables. When I tried
> > to relink the table in Access it complained
> > it couldn't find the table because of a naming violation.
>
> Sounds like psqlODBC may not yet know it's supposed to
> ignore pg_attribute entries with attisdropped true.

Hmm isn't it your responsibility to change the code ?
It was you who rejected all the proposals which don't
need the attisdropped test.

regards,
Hiroshi Inoue
    http://w2422.nsk.ne.jp/~inoue/

Hiroshi Inoue <Inoue@tpf.co.jp> writes:
> Tom Lane wrote:
>> Sounds like psqlODBC may not yet know it's supposed to
>> ignore pg_attribute entries with attisdropped true.

> Hmm isn't it your responsibility to change the code ?

Well, perhaps so, but I have no ability to test ODBC, and it's not
even in the PG CVS tree anymore, so I have no ability to make the
change anyway.

            regards, tom lane