grep for unique values

Recently I output a large chunk of very similar data to a text file. Here is a sample chunk:

--message: 1
type name: SYS.LCR$_ROW_RECORD
source database: DATABASE.DOMAIN.COM
owner: OWNER
object: QRTZ_SCHEDULER_STATE
is tag null: Y
command_type: DELETE
old(1): INSTANCE_NAME
server.domain.com1254958048243
old(2): LAST_CHECKIN_TIME
1254958956909
old(3): CHECKIN_INTERVAL
7500
old(4): RECOVERER
--message: 2
type name: SYS.LCR$_ROW_RECORD
source database: DATABASE.DOMAIN.COM
owner: OWNER
object: QRTZ_SCHEDULER_STATE
is tag null: Y
command_type: INSERT
new(1): INSTANCE_NAME
server.domain.com1254958048243
new(2): LAST_CHECKIN_TIME
1254958964415
new(3): CHECKIN_INTERVAL
7500
new(4): RECOVERER

PL/SQL procedure successfully completed.

This is output from the PRINT_TRANSACTION PL/SQL procedure that Oracle supplies to enable you to troubleshoot Streams Apply errors (from the DBA_APPLY_ERROR view).

We had a large number of errors, but they all seemed to revolve around a few specific objects.

I output the data from PRINT_TRANSACTION for every entry in the DBA_APPLY_ERROR view to a text file. Then used the following grep command to catch exactly which objects were being affected:

grep "object" filename*|sort -t: -k2 -u

Worked a treat :-)

Cheers,

Ian