Data Pump and LOB Counts

Question

I was recently tasked with copying three schemas from one 10gR2 database to another 10gR2 database on the same RHEL server. I decided to use Data Pump in network mode.
After performing the transfer, the count of database objects for the schemas were the same EXCEPT for LOBs. There were significantly fewer on the target (new) database as opposed to the source (old) database.
To be certain, I retried the operation using an intermediate dump file. Again, everything ran without error. The counts were the same as before - fewer LOBs. I also compared row counts of the tables, and they were identical on both systems.
Testing by the application user seemed to indicate everything was fine.
My assumption is that consolidation is going on when the data is imported, resulting in fewer LOBs ... haven't worked much (really, at all) with them before. Just looking for confirmation that this is the case and that nothing is "missing".
Here are the results:
ORIGINAL SOURCE DATABASE
OWNER                          OBJECT_TYPE         STATUS         CNT
COGAUDIT_DEV                   INDEX               VALID            6
                               LOB                 VALID           12
                               TABLE               VALID           21
COGSTORE_DEV                   INDEX               VALID          286
                               LOB                 VALID          390
                               SEQUENCE            VALID            1
                               TABLE               VALID          200
                               TRIGGER             VALID            2
                               VIEW                VALID            4
PLANNING_DEV                   INDEX               VALID           37
                               LOB                 VALID           15
                               SEQUENCE            VALID            3
                               TABLE               VALID           31
                               TRIGGER             VALID            3
14 rows selected.Here are the counts on the BOPTBI (target) database:
NEW TARGET DATABASE
OWNER                          OBJECT_TYPE         STATUS         CNT
COGAUDIT_DEV                   INDEX               VALID            6
                               LOB                 VALID            6
                               TABLE               VALID           21
COGSTORE_DEV                   INDEX               VALID          286
                               LOB                 VALID           98
                               SEQUENCE            VALID            1
                               TABLE               VALID          200
                               TRIGGER             VALID            2
                               VIEW                VALID            4
PLANNING_DEV                   INDEX               VALID           37
                               LOB                 VALID           15
                               SEQUENCE            VALID            3
                               TABLE               VALID           31
                               TRIGGER             VALID            3
14 rows selected.We're just curious ... thanks for any insight on this!
Chris
Edited by: 877086 on Aug 3, 2011 4:38 PM
Edited by: 877086 on Aug 3, 2011 4:40 PM
Edited by: 877086 on Aug 3, 2011 4:40 PM

Answer 1

OK, here is the SQL that produced the object listing:
break on owner skip 1;
select   owner,
         object_type,
         status,
         count(*) cnt
from     dba_objects
where    owner in ('COGAUDIT_DEV', 'COGSTORE_DEV', 'PLANNING_DEV')
group by owner,
         object_type,
         status
order by 1, 2;Here is the export parameter file:
# cog_all_exp.par
userid = chamilton
content = all
directory = xfer
dumpfile = cog_all_%U.dmp
full = n
job_name = cog_all_exp
logfile = cog_all_exp.log
parallel = 2
schemas = (cogaudit_dev, cogstore_dev, planning_dev)Here is the import parameter file:
# cog_all_imp.par
userid = chamilton
content = all
directory = xfer
dumpfile = cog_all_%U.dmp
full = n
job_name = cog_all_imp
logfile = cog_all_imp.log
parallel = 2
reuse_datafiles = n
schemas = (cogaudit_dev, cogstore_dev, planning_dev)
skip_unusable_indexes = n
table_exists_action = replaceThe above parameter files were for the dumpfile version. For the original network link version, I omitted the dumpfile parameter and substituted "network_link = boptcog_xfer".
Chris
Edited by: 877086 on Aug 3, 2011 6:18 PM