Thread: Long term read-only records?

Is there a way to make a record read-only for months or years? My intent is
that after a cutoff (a program can run making rows r/o) existing records can
no longer be modified without unlocking first. User security is
insufficient, since that only applies to tables. I'd rather not have to
maintain a R/O flag (or date) in code, because it relies on correct,
voluntary implementation every time. I'd like something more enforceable
then that, and that has me looking to the DBMS

Thank you for your time.

You can also set users to not do things like triggers, views, functions, etc,
and basically leave nothing but SELECT. That won't work?

Jason Hihn wrote:
> Is there a way to make a record read-only for months or years? My intent is
> that after a cutoff (a program can run making rows r/o) existing records can
> no longer be modified without unlocking first. User security is
> insufficient, since that only applies to tables. I'd rather not have to
> maintain a R/O flag (or date) in code, because it relies on correct,
> voluntary implementation every time. I'd like something more enforceable
> then that, and that has me looking to the DBMS
>
> Thank you for your time.
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 4: Don't 'kill -9' the postmaster
>

OOOPS,
    record by record, I see. Mabye the flag in combination with a view?

Dennis Gearon wrote:
> You can also set users to not do things like triggers, views, functions,
> etc, and basically leave nothing but SELECT. That won't work?
>
> Jason Hihn wrote:
>
>> Is there a way to make a record read-only for months or years? My
>> intent is
>> that after a cutoff (a program can run making rows r/o) existing
>> records can
>> no longer be modified without unlocking first. User security is
>> insufficient, since that only applies to tables. I'd rather not have to
>> maintain a R/O flag (or date) in code, because it relies on correct,
>> voluntary implementation every time. I'd like something more enforceable
>> then that, and that has me looking to the DBMS
>>
>> Thank you for your time.
>>
>>
>> ---------------------------(end of broadcast)---------------------------
>> TIP 4: Don't 'kill -9' the postmaster
>>
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org
>

On Wednesday 26 Mar 2003 5:09 pm, Dennis Gearon wrote:
> OOOPS,
>     record by record, I see. Mabye the flag in combination with a view?

That's what I'd do. Don't forget to make sure you still have a user who can
delete the underlying r/o records at some point in the future.
--
  Richard Huxton

On Wed, 2003-03-26 at 11:34, Jason Hihn wrote:
> Is there a way to make a record read-only for months or years? My intent is
> that after a cutoff (a program can run making rows r/o) existing records can
> no longer be modified without unlocking first. User security is
> insufficient, since that only applies to tables. I'd rather not have to
> maintain a R/O flag (or date) in code, because it relies on correct,
> voluntary implementation every time. I'd like something more enforceable
> then that, and that has me looking to the DBMS

Create a rule or trigger on the table to check the age of a date column
and then either disregard or allow updates.

Robert Treat

Thank you all for your help. I've been working with MySQL too long. ;-)

> -----Original Message-----
> From: Robert Treat [mailto:xzilla@users.sourceforge.net]
> Sent: Thursday, March 27, 2003 3:18 PM
> To: Jason Hihn
> Cc: pgsql-general@postgresql.org
> Subject: Re: [GENERAL] Long term read-only records?
>
>
> On Wed, 2003-03-26 at 11:34, Jason Hihn wrote:
> > Is there a way to make a record read-only for months or years?
> My intent is
> > that after a cutoff (a program can run making rows r/o)
> existing records can
> > no longer be modified without unlocking first. User security is
> > insufficient, since that only applies to tables. I'd rather not have to
> > maintain a R/O flag (or date) in code, because it relies on correct,
> > voluntary implementation every time. I'd like something more enforceable
> > then that, and that has me looking to the DBMS
>
> Create a rule or trigger on the table to check the age of a date column
> and then either disregard or allow updates.
>
> Robert Treat
>
>