Seems like a Painful Incompleteness Awkwardness (PIA), also known shorter and
more colorfully.
Still, I wonder if you have a pause on DML whlist you are cloning the schema.
IF you do, a woeful work-around might be to temporarily shut off the when NULL
sequence directive on the table(s) in THOMAS, do the clone, turn them back on
on THOMAS and create the when NULL sequence directive on ARTHUR.
Presumably it is only the when NULL sequence directive that is Fouled Up Beyond
All Recognition (FUBAR) and the sequence itself is getting correctly created
with the current next number. Right?
What a PIA!
mwf
PS: Another way would be to create a new pluggable database and carefully
import THOMAS from the current location into the new pluggable. Some might say
that is a better testbed, unless of course you are testing the rename schema
completeness of clone schema!
-----Original Message-----
From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx] On ;
Behalf Of Mladen Gogala
Sent: Monday, July 26, 2021 7:40 PM
To: oracle-l@xxxxxxxxxxxxx
Subject: Re: datapump import schema and default values taken from a sequence
Nope. If REMAP_SCHEMA doesn't work, that's it. You can, however, pre-create the
with the proper definition and then use TABLE_EXISTS_ACTION=APPEND. That,
however, could lead to another problem since the newly created sequence will
start from the beginning, whatever START WITH option of the CREATE SEQUENCE
command has specified, so the rows from the original table will very likely
differ from the imported rows. Let me discover the hot water and tell you that
you're in a pickle. Good luck.
On 7/26/21 6:03 AM, Thomas Kellerer (Redacted sender thomas.kellerer for
DMARC) wrote:
So my question is:
Is there any way to tell DataPump to replace the reference to
"THOMAS"."MY_TABLE_ID_SEQ" with the target schema:
"ARTHUR"."MY_TABLE_ID_SEQ"?
