doing import with code that make us of dbms_datapump procedes and also with impdp utility. Three scenarios to meet, it can be done separately. right now its dead slow as parallelism not used. its exadatabase.
1 source table is non partitioned , target table into which data to imported is partitioned, irrespective of parallelism, only one worker process doing all the job and it is dead slow and running for days. we dont see records in select count(*). lot of unod,redos.
48 dump files. how do I make import run in parallel? although it shows rows imported in worker process, i can not query them?. worker process are not getting created even giving parallelism=48
dumpfile=%U.dmp
logfile=.log
parallel=48
directory=
include=TABLE:"IN ('TABLE_NAME1','TABLE_NAME2)"
schemas=PRODREL
DATA_OPTIONS=TRUST_EXISTING_TABLE_PARTITIONS
METRICS=Y
LOGTIME=ALL
CONTENT=DATA_ONLY
table_exists_action=APPEND
DATA_OPTIONS=SKIP_CONSTRAINT_ERRORS
Worker 1 Status:
Instance ID: 1
Process Name: DW06
State: EXECUTING
Object Schema:
Object Name: TABLE_NAME1
Object Type: SCHEMA_EXPORT/TABLE/TABLE_DATA
Completed Objects: 1
Completed Rows: 432,929,844
Completed Bytes:
Percent Done: 4
Worker Parallelism: 1
Import>
2. partition table dump , into another partition table that has different partition strategy (different partition key) but has same table DDL (Same number of columns, same datatypes etc., data only import.. lowest source table partition data that have no equvilent partitions in target partition table should go to first partition of target partition table instead of skipping whole partition altogether. using same as shown above.
wait events seen: wait for unread message on broadcast channel,PX Deq: Execution Msg , PX Deq: Execute Reply -its time same as impdp duration time.
3. non partition table import with QUERY clause into partition table. there are several millions of records which are getting rejected one row at a time as per the imp log file. no parallelism being used.. ..
any help please on these three cases. thanks in advance. what I am looking is to perform import in faster method in existing tables that has identical number of column+same datatypes etc. and in parallel or any other best way of doing it.