On Fri, 27 Jun 2008 07:39:27 +0100, "Roy Hann"
<specially@[EMAIL PROTECTED]
> wrote:
>"Paul White" <pwhite@[EMAIL PROTECTED]
> wrote in message
>news:mailman.302.1214520577.2607.info-ingres@[EMAIL PROTECTED]
>> Anyone care to comment on this thread at openroad-users?
>> I'm sure it must have been covered at cdi in the past.
>> I reckon copydb (also unloaddb?) needs a "schema only" type option
>
>I daresay there are such scripts around and I have written (and now lost)
>such things myself. They are a pain to write.
>
>Which is why Chris Dawe gave up his weekend a couple of weeks ago to
enhance
>copydb during the IUA sprint so that you can direct it to produce just
the
>bits you want. Obviously that's not yet released, but it is coming soon.
>And if you want to do your own build of copydb there's nothing to stop
you.
>I don't know the best way to contact Chris this week but I'm sure he'll
see
>this and join in.
>
>Roy
Ingres 2006r2 already has '-with_index(es)?' '-with_modify',
'-with_permits', etc. I had previously used therm to generate
overnight housekeeping scripts. For each table:
copydb DBNAME TABLENAME -with_modify -with_index -infile=TABLENAME.mod
What we added were the negatives (e.g. -all -nopermits) and we added
all these flags to unloaddb.
We also make ASCII unloads and copys (with '-c') specify the
II_DECIMAL, II_DATE_FORMAT and II_MONEY_FORMAT at the top of the
copy.in script.
All these are with Ingres for QA and formal release, together with
some extra comments from after the Sprint. We only set the FORMAT
values if they were specified (as Ingres or OS variables). With
hindsight, it would e best to ALWAYS specify them, so if an American
(with all default values) sent an ASCII unload to someone in Germany
(with all three set to non-default values), the copy.in needs to
specify the American defaults to get the database to load cleanly
first time.
Chris
--
email address available from http://www.iua.org.uk/committee.htm
--
uSsPeAnMet@[EMAIL PROTECTED]
all SPAM to reply
|
