If, however, you are using the
online catalog from the current database, you can only analyze redo log files from the current
database.
Since you can run LogMiner from one database against the redo log files in another database,
the character sets used on both databases must match. The hardware platform must also match
the one used when the redo log files were generated.
Extracting the Data Dictionary
One potential problem with extracting the data dictionary to a flat file is that while you are
extracting the data dictionary, someone else could be issuing DDL statements. Therefore, the
extracted data dictionary could be out of sync with the database. When you use a flat file to
store the data dictionary, fewer system resources are required than when you use redo log files.
When you extract the data dictionary to redo log files, no DDL statements can be processed
during the time in which the data dictionary is extracted. Therefore, the dictionary will be in sync
with the database; the extraction is more resource intensive, but the extraction process is faster.
To extract the data dictionary to either a flat file or to redo log files, you use the procedure
DBMS_LOGMNR_D.BUILD. The data dictionary file is placed in a directory. Therefore, you must
have write permission for the directory in which the file will be placed. To define the location of
the directory, use the initialization parameter UTL_FILE_DIR. For example, to specify the location
D:\Oracle\Ora10\database as the location for the LogMiner output, you place the following
entry in the parameter file:
UTL_FILE_DIR= D:\Oracle\Ora10\database
NOTE
You cannot dynamically change the UTL_FILE_DIR parameter using
the alter system command.
Pages:
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759